CN111539894A - Novel image enhancement method - Google Patents
Novel image enhancement method Download PDFInfo
- Publication number
- CN111539894A CN111539894A CN202010351896.3A CN202010351896A CN111539894A CN 111539894 A CN111539894 A CN 111539894A CN 202010351896 A CN202010351896 A CN 202010351896A CN 111539894 A CN111539894 A CN 111539894A
- Authority
- CN
- China
- Prior art keywords
- image
- information
- value
- formula
- quantum
- 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
- 238000000034 method Methods 0.000 title claims abstract description 52
- 230000008569 process Effects 0.000 claims abstract description 20
- 238000012545 processing Methods 0.000 claims abstract description 20
- 238000010606 normalization Methods 0.000 claims description 15
- 230000000694 effects Effects 0.000 claims description 12
- 238000000354 decomposition reaction Methods 0.000 claims description 9
- 238000002474 experimental method Methods 0.000 claims description 9
- 238000005070 sampling Methods 0.000 claims description 6
- 230000009466 transformation Effects 0.000 claims description 6
- 230000005283 ground state Effects 0.000 claims description 5
- 230000002708 enhancing effect Effects 0.000 claims description 4
- 238000001914 filtration Methods 0.000 claims description 4
- 239000003086 colorant Substances 0.000 claims description 3
- 230000008602 contraction Effects 0.000 claims description 3
- 238000013519 translation Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 description 10
- 238000005259 measurement Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000005610 quantum mechanics Effects 0.000 description 2
- 238000009795 derivation Methods 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000000873 masking effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N10/00—Quantum computing, i.e. information processing based on quantum-mechanical phenomena
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/10—Image enhancement or restoration using non-spatial domain filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20048—Transform domain processing
- G06T2207/20064—Wavelet transform [DWT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
Landscapes
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Mathematical Analysis (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Condensed Matter Physics & Semiconductors (AREA)
- Computational Mathematics (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Image Processing (AREA)
Abstract
The invention provides a novel image enhancement method. The novel image enhancement method comprises the following steps: s1: inputting an image; s2: expressing the image in a quantum mode; s3: collecting an image; s4: initializing an image gray threshold T and an image information entropy E; s5: normalizing the image; s6: image enhancement processing; s7: calculating the entropy of the image information; s8: judging whether the information entropy is maximum or not; if not, returning to the step S4; if yes, the process proceeds to the next step S9; s9: normalizing the T value corresponding to the maximum information entropy; and S10, outputting the image. The novel image enhancement method provided by the invention has the advantages of good detail information storage, no phenomena of sharpness and the like, and obvious target display in the image.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a novel image enhancement method.
Background
The quantum derivation thought is inspired by algorithms derived from natural laws such as artificial neural networks, genetic algorithms and the like, and then existing algorithms are improved or new algorithms are developed by utilizing concepts and principles of quantum mechanics and a common physical and chemical mathematical system and are applied to classical computers. In 1997, the concept of quantum image processing was first proposed by Vlasov, and in 2003, Beach and Venegas-Andraca gave their own quantum image processing algorithms, respectively, and attempted to apply the existing algorithm (Grover quantum search algorithm) to images, and quantum image processing has formally started to be of interest. In 2005, Latore proposed a new quantum image representation method, and in 2006, Google first utilized a quantum neural network learning method to reduce the error rate of speech recognition by 20% -30%, and the error rate of image recognition from 26% to 15%. In 2007, Schekov et al proposed quantum-derived morphological edge detection based on theories and concepts such as quantum mechanics and quantum information. In 2010, Wei Pao et al propose a quantum statistical probability-based medical image enhancement algorithm, which combines the characteristics of medical images to construct quantum enhancement operators in 3 x 3 windows. In 2011, people such as high still waves apply a quantum enhancement algorithm to a color image, the algorithm firstly converts an RGB color model of the image into an HSV model, performs quantum enhancement on a luminance component, and finally converts the enhanced HSV image model into an RGB color image. After 2010, research on quantum image processing algorithms is gradually flourishing, the effect is remarkable, and the application fields are increased year by year.
Remote sensing image enhancement is a process for preprocessing a remote sensing image, and the enhancement effect of the remote sensing image enhancement plays a crucial role in later image segmentation and image classification and identification. Image enhancement techniques can be divided into spatial domain techniques and transform domain techniques. The spatial domain is directly operated on the pixel points and the pixel values of the image. Classical methods are histogram equalization and its improvement, linear stretching, unsharp masking, etc. Histogram equalization is a global image enhancement method, ignores local information, and cannot maintain the average brightness level of an image, which may result in over-saturation or under-saturation of a selected region. The unsharp mask can enhance image noise in the process of enhancing image details. The method of converting an image to the frequency domain and processing the converted coefficients is called transform domain processing. The classic remote sensing image frequency domain enhancement algorithm comprises image enhancement based on wavelet transformation, image enhancement based on NSSCT, homomorphic filtering and the like. These methods of transforming the domain are superior to the spatial domain, but take longer and have more complex algorithms. In order to solve these problems and improve the quality of the remote sensing image, it is necessary to improve the existing algorithm or propose a new algorithm.
Disclosure of Invention
The invention aims to provide a novel image enhancement method which has the advantages of good detail information storage, no sharp phenomenon and obvious target display in a picture.
In order to solve the above technical problem, the novel image enhancement method provided by the present invention comprises the following steps:
s1: inputting an image;
s2: expressing the image in a quantum mode;
s3: collecting an image;
s4: initializing an image gray threshold T and an image information entropy E;
s5: normalizing the image;
s6: image enhancement processing;
s7: calculating the entropy of the image information;
s8: judging whether the information entropy is maximum or not; if not, returning to the step S4; when judging
Yes, proceed to next step S9;
s9: normalizing the T value corresponding to the maximum information entropy;
and S10, outputting the image.
Preferably, the quantum-mode representation image in S2 includes the following steps:
assuming a quantum system comprising two quanta, consisting of states and states, four ground states are generated, namely:
in the formula (1-1)The representation is a tensor product, also called product of values and kronecker product. These four ground states may exist with some probability simultaneously in a two quantum system.
While a two-dimensional (2D) image with resolution H x W requires a number of quantum bits of log2H+log2W + Q, Q represents the color depth of the image:
a two-dimensional (2D) image can be represented as:
in the formula | XY>Is represented by the coordinate information of the image pixel, | CXY>Representing color information, the formula is as follows:
the above shows a grayscale image, but also an RGB color image, i.e. when Q is 24, the following formula shows three primary colors of red, green and blue:
preferably, the steps S1, S2 and S3 include the following steps:
setting an initial image as g (m, n), inputting the initial image g (m, n), carrying out quantum expression coding by adopting formulas (1-7) to (1-10), carrying out expression digital image, and sampling the expression image to obtain a sampling image f (m, n).
Preferably, the steps S4 and S5 are as follows:
by the following formulas (2-1) (2-2),
when g (m, n) ≦ T:
when g (m, n) ≧ T:
the method comprises the following steps that g (m, n) represents an original image, f (m, n) represents an image after normalization processing, the value range of a pixel value is [0,1], min and max are respectively the minimum value and the maximum value of the original image pixel value, the coefficient lambda belongs to [0,1], the optimal values of different image normalization coefficients lambda are different, experiments show that lambda is 0.5 and is suitable for remote sensing images, T is a normalization threshold, a threshold T is determined through image information entropy maximization, after the maximum information entropy is determined, the corresponding threshold can be determined, and the image information entropy formula is as follows:
selecting a neighborhood gray mean value of an image as a spatial feature quantity of gray distribution, forming a feature binary group with pixel gray of the image, and recording the feature binary group as (i, j), wherein i represents a gray value (0< ═ i < ═ 255) of a pixel, and j represents a neighborhood gray (0< ═ j < > 255), the above formula can reflect the comprehensive feature of the gray value at a certain pixel position and the gray distribution of surrounding pixels, wherein f (i, j) is the frequency of the feature binary group (i, j), and N is the scale of the image, and performing threshold normalization processing on the sampled image g (m, N) through the above formula.
Preferably, the steps S6 and S7 are as follows:
the normalized image is enhanced, and whether the entropy of the image information is the maximum value is judged, at this time, the wavelet transform decomposition operation is needed to be carried out on the image: decomposing the image into a high-frequency information part and a low-frequency information part, carrying out nonlinear transformation on the obtained wavelet coefficient part, enhancing the high-frequency detail information part of the image, carrying out wavelet coefficient filtering operation by adopting a current threshold T, and reconstructing the wavelet coefficient to obtain an enhanced image by using a formula (2-4) to a formula (2-8);
the one-dimensional wavelet transform is a series expansion of a function, and is formed by a set of phi (t), psi (t) through expansion and contraction translation, and the formula is as follows:
wherein phi (t-k), (2)j/2ψ(2jt-k)), (k ═ infinity. + ∞, j ═ 0. + ∞) is a pair of orthogonal bases, c ∞kdj,kIs the inner product of the transformed function and the orthogonal basis:
phi (t), psi (t) are expressed as a scale function and a wavelet function, respectively, satisfying the following equations:
passing noisy information through a low-pass filter h0And a high-pass filter h1And then is broken down into two parts: a low-frequency information part and a high-frequency noise part, wherein the process is called one-layer wavelet transformation;
the wavelet reconstruction process is the inverse of the decomposition process: in the Mallat algorithm reconstruction process, 0 is inserted into the separated points of the decomposed signal firstly, because the separated points in the previous decomposition process are sampled and are called to be extended, convolution is carried out on the separated points and the filters respectively (low-frequency information is convolved with a low-pass filter, high-frequency information is convolved with a high-pass filter), the obtained results are added, and finally, the extended part is removed, so that the image information can be obtained.
Preferably, the specific steps of step S8 are as follows:
and (4) judging whether the image information entropy in the step (S7) is the maximum, if not, continuously modifying the threshold value until the maximum information entropy value is obtained, repeating the steps (S4) to (S7), and when the obtained image reaches the maximum information entropy, wherein the corresponding threshold value T is the threshold value with the best enhancement effect.
Preferably, the steps S9 and S10 are as follows:
according to the threshold value of the optimal enhancement effect, threshold value normalization processing is carried out on the original image g (m, n) by using formulas (2-1) and (2-2) to obtain a normalized image f (m, n), and the enhanced image is output, namely the image after quantum enhancement.
Compared with the related art, the novel image enhancement method provided by the invention has the following beneficial effects:
compared with other image enhancement algorithms, the quantum image enhancement algorithm provided by the invention can generally improve the information entropy index by more than 7%, the image definition index by more than 14%, the image quality measurement CE index by more than 5%, the detail information of the image is stored more perfectly, the phenomena of sharpness and the like are avoided, the target display in the image is more obvious, and the overall definition of the image is improved.
Drawings
FIG. 1 is a flow chart of a novel image enhancement method provided by the present invention;
FIG. 2 is a diagram of an experimental image effect of a remote sensing image;
fig. 3 is an image effect diagram of a human image experiment.
Detailed Description
The invention is further described with reference to the following figures and embodiments.
Referring to fig. 1-3, the novel image enhancement method includes the following steps:
s1: inputting an image;
s2: expressing the image in a quantum mode;
s3: collecting an image;
s4: initializing an image gray threshold T and an image information entropy E;
s5: normalizing the image;
s6: image enhancement processing;
s7: calculating the entropy of the image information;
s8: judging whether the information entropy is maximum or not; if not, returning to the step S4; when judging
Yes, proceed to next step S9;
s9: normalizing the T value corresponding to the maximum information entropy;
and S10, outputting the image.
The quantum mode expression image in the S2 comprises the following steps:
assuming a quantum system comprising two quanta, consisting of states and states, four ground states are generated, namely:
in the formula (1-1)The representation is a tensor product, also called product of values and kronecker product. The four ground states are two in numberThere may be some probability of simultaneous existence within a subsystem.
While a two-dimensional (2D) image with resolution H x W requires a number of quantum bits of log2H+log2W + Q, Q represents the color depth of the image:
a two-dimensional (2D) image can be represented as:
in the formula | XY>Is represented by the coordinate information of the image pixel, | CXY>Representing color information, the formula is as follows:
the above shows a grayscale image, but also an RGB color image, i.e. when Q is 24, the following formula shows three primary colors of red, green and blue:
the specific steps of the step S1, the step S2 and the step S3 are as follows:
setting an initial image as g (m, n), inputting the initial image g (m, n), carrying out quantum expression coding by adopting formulas (1-7) to (1-10), carrying out expression digital image, and sampling the expression image to obtain a sampling image f (m, n).
The specific steps of step S4 and step S5 are as follows:
by the following formula (2-1) (2-2)
When g (m, n) ≦ T:
when g (m, n) ≧ T:
the method comprises the following steps that g (m, n) represents an original image, f (m, n) represents an image after normalization processing, the value range of a pixel value is [0,1], min and max are respectively the minimum value and the maximum value of the original image pixel value, the coefficient lambda belongs to [0,1], the optimal values of different image normalization coefficients lambda are different, experiments show that lambda is 0.5 and is suitable for remote sensing images, T is a normalization threshold, a threshold T is determined through image information entropy maximization, after the maximum information entropy is determined, the corresponding threshold can be determined, and the image information entropy formula is as follows:
selecting a neighborhood gray mean value of an image as a spatial feature quantity of gray distribution, forming a feature binary group with pixel gray of the image, and recording the feature binary group as (i, j), wherein i represents a gray value (0< ═ i < ═ 255) of a pixel, and j represents a neighborhood gray (0< ═ j < > 255), the above formula can reflect the comprehensive feature of the gray value at a certain pixel position and the gray distribution of surrounding pixels, wherein f (i, j) is the frequency of the feature binary group (i, j), and N is the scale of the image, and performing threshold normalization processing on the sampled image g (m, N) through the above formula.
The specific steps of step S6 and step S7 are as follows:
the normalized image is enhanced, and whether the entropy of the image information is the maximum value is judged, at this time, the wavelet transform decomposition operation is needed to be carried out on the image: decomposing the image into a high-frequency information part and a low-frequency information part, carrying out nonlinear transformation on the obtained wavelet coefficient part, enhancing the high-frequency detail information part of the image, carrying out wavelet coefficient filtering operation by adopting a current threshold T, and reconstructing the wavelet coefficient to obtain an enhanced image by using a formula (2-4) to a formula (2-8);
the one-dimensional wavelet transform is a series expansion of a function, and is formed by a set of phi (t), psi (t) through expansion and contraction translation, and the formula is as follows:
wherein phi (t-k), (2)j/2ψ(2jt-k)), (k ═ infinity. + ∞, j ═ 0. + ∞) is a pair of orthogonal bases, c ∞kdj,kIs the inner product of the transformed function and the orthogonal basis:
phi (t), psi (t) are expressed as a scale function and a wavelet function, respectively, satisfying the following equations:
passing noisy information through a low-pass filter h0And a high-pass filter h1And then is broken down into two parts: a low-frequency information part and a high-frequency noise part, wherein the process is called one-layer wavelet transformation;
the wavelet reconstruction process is the inverse of the decomposition process: in the Mallat algorithm reconstruction process, 0 is inserted into the separated points of the decomposed signal firstly, because the separated points in the previous decomposition process are sampled and are called to be extended, convolution is carried out on the separated points and the filters respectively (low-frequency information is convolved with a low-pass filter, high-frequency information is convolved with a high-pass filter), the obtained results are added, and finally, the extended part is removed, so that the image information can be obtained.
The specific steps of step S8 are as follows:
and (4) judging whether the image information entropy in the step (S7) is the maximum, if not, continuously modifying the threshold value until the maximum information entropy value is obtained, repeating the steps (S4) to (S7), and when the obtained image reaches the maximum information entropy, wherein the corresponding threshold value T is the threshold value with the best enhancement effect.
The specific steps of step S9 and step 10 are as follows:
according to the threshold value of the optimal enhancement effect, threshold value normalization processing is carried out on the original image g (m, n) by using formulas (2-1) and (2-2) to obtain a normalized image f (m, n), and the enhanced image is output, namely the image after quantum enhancement.
The following specific experiments are used to specifically explain the novel image enhancement method provided by the present invention:
experiment one: remote sensing image experiment
The experimental effect is shown in fig. 2;
experimental data for different algorithms are compared as shown in table 1.
Table 1:
the experimental results are as follows: from the experimental results and data of fig. 2 and table 1, it can be derived: compared with other image enhancement algorithms, the quantum image enhancement algorithm provided by the patent can generally improve the information entropy index by more than 10%, can improve the image definition index by more than 14%, and generally improves the image quality measurement function (CE) index by more than 10%. The detail information of the image is stored more perfectly, the phenomena of sharpness and the like can not occur, the target display in the image is more obvious, and the integral definition of the image is improved.
Experiment two: human image experiment
The experimental effect is shown in fig. 3;
experimental data for different algorithms are compared as shown in table 2.
Table 2:
the experimental results are as follows: from the experimental results and data of fig. 3 and table 2, it can be derived: compared with other image enhancement algorithms, the quantum image enhancement algorithm provided by the patent can generally improve the information entropy index by more than 7%, can improve the image definition index by more than 14%, and generally improves the image quality measurement function (CE) index by more than 5%.
Compared with the related art, the novel image enhancement method provided by the invention has the following beneficial effects:
compared with other image enhancement algorithms, the quantum image enhancement algorithm provided by the invention can generally improve the information entropy index by more than 7%, the image definition index by more than 14%, the image quality measurement CE index by more than 5%, the detail information of the image is stored more perfectly, the phenomena of sharpness and the like are avoided, the target display in the image is more obvious, and the overall definition of the image is improved.
The above description is only an embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes, which are made by using the contents of the present specification and the accompanying drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (7)
1. A novel image enhancement method is characterized by comprising the following steps:
s1: inputting an image;
s2: expressing the image in a quantum mode;
s3: collecting an image;
s4: initializing an image gray threshold T and an image information entropy E;
s5: normalizing the image;
s6: image enhancement processing;
s7: calculating the entropy of the image information;
s8: judging whether the information entropy is maximum or not; if not, returning to the step S4; if yes, the process proceeds to the next step S9;
s9: normalizing the T value corresponding to the maximum information entropy;
and S10, outputting the image.
2. The novel image enhancement method according to claim 1, wherein the quantum-wise expressing image in S2 includes the steps of:
assuming a quantum system comprising two quanta, consisting of states and states, four ground states are generated, namely:
in the formula (1-1)The representation is a tensor product, also called the product of values and the kronecker product, which can exist simultaneously with a certain probability in a two-quantum system.
Whereas a two-dimensional (2D) image with resolution H W is requiredThe number of quantum bits is log2H+log2W + Q, Q represents the color depth of the image:
a two-dimensional (2D) image can be represented as:
in the formula | XY>Is represented by the coordinate information of the image pixel, | CXY>Representing color information, the formula is as follows:
|XY>=|X>|Y>=|X0X1...Xw-1>|Y0Y1...Yh-1>(1-8)
the above shows a grayscale image, but also an RGB color image, i.e. when Q is 24, the following formula shows three primary colors of red, green and blue:
3. the novel image enhancement method according to claim 2, wherein the steps S1, S2 and S3 are as follows:
setting an initial image as g (m, n), inputting the initial image g (m, n), carrying out quantum expression coding by adopting formulas (1-7) to (1-10), carrying out expression digital image, and sampling the expression image to obtain a sampling image f (m, n).
4. The novel image enhancement method according to claim 3, wherein the steps S4 and S5 are as follows:
by the following formula (2-1) (2-2)
When g (m, n) ≦ T:
when g (m, n) ≧ T:
the method comprises the following steps that g (m, n) represents an original image, f (m, n) represents an image after normalization processing, the value range of a pixel value is [0,1], min and max are respectively the minimum value and the maximum value of the original image pixel value, the coefficient lambda belongs to [0,1], the optimal values of different image normalization coefficients lambda are different, experiments show that lambda is 0.5 and is suitable for remote sensing images, T is a normalization threshold, a threshold T is determined through image information entropy maximization, after the maximum information entropy is determined, the corresponding threshold can be determined, and the image information entropy formula is as follows:
selecting a neighborhood gray mean value of an image as a spatial feature quantity of gray distribution, forming a feature binary group with pixel gray of the image, and recording the feature binary group as (i, j), wherein i represents a gray value (0< ═ i < ═ 255) of a pixel, and j represents a neighborhood gray (0< ═ j < > 255), the above formula can reflect the comprehensive feature of the gray value at a certain pixel position and the gray distribution of surrounding pixels, wherein f (i, j) is the frequency of the feature binary group (i, j), and N is the scale of the image, and performing threshold normalization processing on the sampled image g (m, N) through the above formula.
5. The novel image enhancement method according to claim 4, wherein the steps S6 and S7 are as follows:
the normalized image is enhanced, and whether the entropy of the image information is the maximum value is judged, at this time, the wavelet transform decomposition operation is needed to be carried out on the image: decomposing the image into a high-frequency information part and a low-frequency information part, carrying out nonlinear transformation on the obtained wavelet coefficient part, enhancing the high-frequency detail information part of the image, carrying out wavelet coefficient filtering operation by adopting a current threshold T, and reconstructing the wavelet coefficient to obtain an enhanced image by using a formula (2-4) to a formula (2-8);
the one-dimensional wavelet transform is a series expansion of a function, and is formed by a set of phi (t), psi (t) through expansion and contraction translation, and the formula is as follows:
wherein phi (t-k), (2)j/2ψ(2jt-k)), (k ═ infinity. + ∞, j ═ 0. + ∞) is a pair of orthogonal bases, c ∞kdj,kIs the inner product of the transformed function and the orthogonal basis:
phi (t), psi (t) are expressed as a scale function and a wavelet function, respectively, satisfying the following equations:
passing noisy information through a low-pass filter h0And a high-pass filter h1And then is broken down into two parts: low frequency information part, high frequency noise part, thisThe process is called one-layer wavelet transform;
the wavelet reconstruction process is the inverse of the decomposition process: in the Mallat algorithm reconstruction process, 0 is inserted into the separated points of the decomposed signal firstly, because the separated points in the previous decomposition process are sampled and are called to be extended, convolution is carried out on the separated points and the filters respectively (low-frequency information is convolved with a low-pass filter, high-frequency information is convolved with a high-pass filter), the obtained results are added, and finally, the extended part is removed, so that the image information can be obtained.
6. The novel image enhancement method according to claim 4, wherein the specific steps of step S8 are as follows:
and (4) judging whether the image information entropy in the step (S7) is the maximum, if not, continuously modifying the threshold value until the maximum information entropy value is obtained, repeating the steps (S4) to (S7), and when the obtained image reaches the maximum information entropy, wherein the corresponding threshold value T is the threshold value with the best enhancement effect.
7. The novel image enhancement method according to claim 4, wherein the steps S9 and S10 are as follows:
according to the threshold value of the optimal enhancement effect, threshold value normalization processing is carried out on the original image g (m, n) by using formulas (2-1) and (2-2) to obtain a normalized image f (m, n), and the enhanced image is output, namely the image after quantum enhancement.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010351896.3A CN111539894A (en) | 2020-04-28 | 2020-04-28 | Novel image enhancement method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010351896.3A CN111539894A (en) | 2020-04-28 | 2020-04-28 | Novel image enhancement method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN111539894A true CN111539894A (en) | 2020-08-14 |
Family
ID=71980266
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010351896.3A Pending CN111539894A (en) | 2020-04-28 | 2020-04-28 | Novel image enhancement method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111539894A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116740343A (en) * | 2022-03-01 | 2023-09-12 | 本源量子计算科技(合肥)股份有限公司 | Image segmentation method and device based on quantum classical mixed neural network |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105184752A (en) * | 2015-09-23 | 2015-12-23 | 成都融创智谷科技有限公司 | Image processing method based on wavelet transform |
CN105303535A (en) * | 2015-11-15 | 2016-02-03 | 中国人民解放军空军航空大学 | Global subdivision pyramid model based on wavelet transformation |
CN106530244A (en) * | 2016-10-26 | 2017-03-22 | 长安大学 | Image enhancement method |
CN106780355A (en) * | 2016-11-17 | 2017-05-31 | 中国科学院遥感与数字地球研究所 | A kind of Enhancement Methods about Satellite Images |
CN107204008A (en) * | 2017-06-08 | 2017-09-26 | 上海海事大学 | Quantum image matching method |
-
2020
- 2020-04-28 CN CN202010351896.3A patent/CN111539894A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105184752A (en) * | 2015-09-23 | 2015-12-23 | 成都融创智谷科技有限公司 | Image processing method based on wavelet transform |
CN105303535A (en) * | 2015-11-15 | 2016-02-03 | 中国人民解放军空军航空大学 | Global subdivision pyramid model based on wavelet transformation |
CN106530244A (en) * | 2016-10-26 | 2017-03-22 | 长安大学 | Image enhancement method |
CN106780355A (en) * | 2016-11-17 | 2017-05-31 | 中国科学院遥感与数字地球研究所 | A kind of Enhancement Methods about Satellite Images |
CN107204008A (en) * | 2017-06-08 | 2017-09-26 | 上海海事大学 | Quantum image matching method |
Non-Patent Citations (1)
Title |
---|
王兵 等: "量子索引图像的描述方法与隐写算法" * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116740343A (en) * | 2022-03-01 | 2023-09-12 | 本源量子计算科技(合肥)股份有限公司 | Image segmentation method and device based on quantum classical mixed neural network |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111583123A (en) | Wavelet transform-based image enhancement algorithm for fusing high-frequency and low-frequency information | |
CN110992292B (en) | Enhanced low-rank sparse decomposition model medical CT image denoising method | |
Kaganami et al. | Optimal Approach for Texture Analysis and Classification based on Wavelet Transform and Neural Network. | |
Dharejo et al. | A deep hybrid neural network for single image dehazing via wavelet transform | |
CN111915518B (en) | Hyperspectral image denoising method based on triple low-rank model | |
CN111815550A (en) | Infrared and visible light image fusion method based on gray level co-occurrence matrix | |
Wang et al. | No-reference stereoscopic image quality assessment using quaternion wavelet transform and heterogeneous ensemble learning | |
CN109003247B (en) | Method for removing color image mixed noise | |
Yang et al. | An effective and comprehensive image super resolution algorithm combined with a novel convolutional neural network and wavelet transform | |
CN109741258B (en) | Image super-resolution method based on reconstruction | |
CN111539894A (en) | Novel image enhancement method | |
CN108090914B (en) | Color image segmentation method based on statistical modeling and pixel classification | |
Alsayyh et al. | A Novel Fused Image Compression Technique Using DFT, DWT, and DCT. | |
CN111275620B (en) | Image super-resolution method based on Stacking integrated learning | |
Pai et al. | Medical color image enhancement using wavelet transform and contrast stretching technique | |
CN115082296B (en) | Image generation method based on wavelet domain image generation frame | |
CN108492264B (en) | Single-frame image fast super-resolution method based on sigmoid transformation | |
CN106780398B (en) | A kind of image de-noising method based on noise prediction | |
Thanh et al. | Single image dehazing with optimal color channels and nonlinear transformation | |
CN113781333A (en) | Method for processing underwater image by GAN network based on guided filtering | |
Pizurica et al. | Wavelet domain denoising of single-band and multiband images adapted to the probability of the presence of features of interest | |
CN115797183B (en) | Image super-resolution reconstruction method | |
Talebi et al. | Asymptotic performance of global denoising | |
Seema et al. | Image contrast enhancement approach using differential evolution and particle swarm optimization | |
CN116777807A (en) | Multisource visual information fusion method and system based on multiscale dictionary |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200814 |
|
RJ01 | Rejection of invention patent application after publication |