CN110148078A

CN110148078A - The two width color image compression encryption methods based on two dimensional compaction perception

Info

Publication number: CN110148078A
Application number: CN201910406019.9A
Authority: CN
Inventors: 柴秀丽; 甘志华; 武海洋; 李征; 毕建强; 支相丞; 王俊; 张苗辉
Original assignee: Henan University
Current assignee: Henan University
Priority date: 2019-05-15
Filing date: 2019-05-15
Publication date: 2019-08-20
Anticipated expiration: 2039-05-15
Also published as: CN110148078B

Abstract

The present invention provides a kind of two width color image compression encryption methods based on two dimensional compaction perception.This method includes two stages: in the first stage, carrying out compression measurement to two width colour plaintext images using two dimensional compaction perception, then carries out scramble, quantization to measured value, obtain compressed ciphertext image.In second stage, color host image is normalized first, two components in the color three-component for the color host image that ciphertext image is embedded into after normalization again, it will be embedded into another color component of color host image to relevant key parameter in plain text simultaneously, then final to obtain visually safe ciphertext image by reconfiguring the color three-component of color host image.The present invention can be realized simultaneously the data safety and appearance safety of two width color images.

Description

The two width color image compression encryption methods based on two dimensional compaction perception

Technical field

The present invention relates to image encryption technology field more particularly to a kind of two width color images based on two dimensional compaction perception Compression encryption method.

Background technique

With the fast development of computer networking technology, Internet user can share multi-medium data (example by finger tip Such as file, audio, image, video) etc., exchange, the shared and storage of image information, network security problem can be seen everywhere It becomes increasingly conspicuous, image encryption is one of hot topic of field of image processing.Since color image includes information abundant, have The features such as redundancy is high, and correlation is strong between adjacent pixel and various dimensions, is widely used in the various fields such as remote sensing, medicine. And in existing resume image, it is much encrypted both for gray level image, in order to protect in transimission and storage Color image safety, there is an urgent need to design the algorithm that safe and efficient encryption is carried out to color image.

In addition, most of resume image is encrypted for single image at present, current big data when For under background, only being encrypted to single image cannot more preferable, faster serving-multimedia user.For this reason, it may be necessary to Design it is a kind of for more sub-pictures especially two images and meanwhile encryption algorithm.The appearance of compressed sensing is so that be performed simultaneously Compression and encryption are possibly realized, and are compared with one-dimensional compressed sensing (Compressive sensing, CS), and two dimensional compaction perception can In the case where acquisition original image less information, original image is reconstructed, memory space is greatly saved, improves operation effect Rate, therefore two dimensional compaction perception (2D CS) is more suitable for applying in image encryption under the background of big data.

Original image is encrypted to the meaningless ciphertext of similar noise by current most of resume images before transmission Image.It is worth noting that, meaningless ciphertext image can cause more concerns of attacker hacker in transmission channel, and to cutting The ciphertext image obtained carries out selective analysis.Therefore, it is necessary to a kind of Image Encryption Scheme for taking into account visual security be designed, with same When guarantee image data safety and visual security.In addition, most of resume image is all the information using plaintext image Such as pixel average, cryptographic Hash, comentropy etc. carry out algorithm for design, to increase the connection between algorithm and plaintext, improve algorithm Anti- plaintext attack.But require to send these parameters relevant to cleartext information to recipient in decryption, when right When large batch of image is encrypted simultaneously, need the information content of additional transmissions bigger, this will will cause serious transmission burden, In addition, these parameters are during transmission, once it is stolen by hacker, it is more likely that plaintext image is cracked out, there are certain Security risk.

Summary of the invention

Visual security larger for memory space present in conventional images Encryption Algorithm, not can guarantee image and The problems such as needing the information content of additional transmissions larger when decryption, the present invention provide a kind of two width cromograms based on two dimensional compaction perception As compression encryption method, the data safety and appearance safety of two width color images can be realized simultaneously.

The present invention provides a kind of two width color image compression encryption methods based on two dimensional compaction perception, this method comprises:

Step 1: extracting the first width colour plaintext image P respectively₁With the second width colour plaintext image P₂Color three-component, It is denoted as first group of color three-component and second group of color three-component, and using two-dimensional discrete wavelet conversion respectively to two groups of colors three Component carries out rarefaction；Wherein, P₁And P₂Size be M × N and M=N；

Step 2: calculating separately to obtain according to two groups of color three-components of two width colour plaintext images colored with two width The relevant first group information parameter of plaintext image and the second group information parameter；

Step 3: generating calculation matrix using LSS chaos system and Kronecker product, and the calculation matrix is carried out excellent Change；

Step 4: two dimensional compaction being carried out to two groups of color three-components after rarefaction respectively using the calculation matrix after optimization Perception obtains corresponding first group of compressed sensing measurement sequence and second group of compressed sensing measurement sequence；

Step 5: shuffle operations being carried out to two groups of compressed sensings measurement sequences respectively using sextuple hyperchaotic system, are obtained pair The first group of Scrambling Matrix and second group of Scrambling Matrix answered, and the member by element value in two groups of Scrambling Matrixes less than preset threshold TS Plain value is changed to 0, obtains revised two groups of Scrambling Matrixes；

Step 6: revised two groups of Scrambling Matrixes being quantified respectively, element value is transformed between 0 to 255, is obtained To corresponding first group of quantization matrix and second group of quantization matrix；

Step 7: the color three-component of the color host image F ' after extracting image normalization is denoted as third group color three and divides Amount；Wherein, the size of color host image F ' is M × N；

Step 8: generating 256 cryptographic Hash using the SHA256 function of the first width colour plaintext image, and by described 256 Every 8 one group of cryptographic Hash is converted into 32 decimal number k₁, k₂..., k₃₂, 32 decimal numbers are denoted as cryptographic Hash K；

Step 9: will the first group information parameter relevant to two width colour plaintext images and the second group information parameter according to setting Fixed Parameter Switch rule is respectively converted into first group of integer set and second group of integer set, wherein first group of integer Set and second group of integer set include L integer；

Step 10: (32+2L) a pixel value before one of component in third group color three-component is replaced with into institute Cryptographic Hash K and 2L integer are stated, the first ciphertext matrix containing key parameter information is obtained；

Step 11: other two component in third group color three-component being subjected to DWT decomposition, and will be obtained after decomposition Matrix carries out piecemeal, obtains corresponding first group of carrier matrix and second group of carrier matrix；

Step 12: first group of quantization matrix and second group of quantization matrix are respectively embedded in two groups of carrier matrixs, Obtain the second ciphertext matrix and third ciphertext matrix containing ciphertext image information；

Step 13: first ciphertext matrix, the second ciphertext matrix and third ciphertext matrix being combined, obtained and two The corresponding visual security image of width colour plaintext image.

Further, this method further include: generate the initial value of sextuple hyperchaotic system and LSS chaos system；Wherein:

The initial value for generating sextuple hyperchaotic system specifically includes:

Step A1: intermediate parameters H is calculated according to formula (9) according to the cryptographic Hash K₁~H₆:

Wherein, (t₁,t₂,t₃,t₄,t₅) ∈ (0,1) be preset-key parameter；sum(k₁₆,k₁₇,…,k₂₀) indicate to seek k₁₆, k₁₇,…,k₂₀Sum；max(k₁₆,k₁₇,…,k₂₀) indicate to seek k₁₆, k₁₇,…,k₂₀Maximum value；Indicate the exclusive or of x and y Operation；

Step A2: according to the intermediate parameters H₁~H₆Wherein at the beginning of three of sextuple hyperchaotic system are calculated according to formula (10) Initial value x₀、y₀And z₀:

Wherein, mod (a, b) indicates a to the modulo operation of b, and abs (x) indicates to seek the absolute value of x, and floor (x) indicates meter Calculate the maximum integer for being not more than x；

Step A3: according to the first group information parameter and the intermediate parameters H₁~H₆First group is calculated according to formula (11) Intermediate key r₁、g₁And b₁:

Wherein, Q_R、Q_GAnd Q_BFor the first group information parameter；

Step A4: according to the intermediate key r₁、g₁And b₁And intermediate parameters H₄~H₆It is calculated according to formula (12) sextuple super The other three initial value u of chaos system₀、v₀And w₀:

The initial value for generating LSS chaos system specifically includes:

Step B1: according to the second group information parameter, the intermediate parameters H₁~H₆With preset-key parameter t₁~t₅It presses Illuminated (13) calculates second group of intermediate key r '₁、g′₁With b '₁:

Wherein, Q '_R、Q′_GWith Q '_BFor the second group information parameter；

Step B2: according to second group of intermediate key r '₁、g′₁With b '₁The first of LSS chaos system is calculated according to formula (14) Initial value a₀With parameter r ':

Wherein, (t₆,t₇) ∈ (0,1) be preset-key parameter.

Further, the step 2 specifically:

The first group information parameter and the second group information parameter are calculated according to formula (8):

Wherein, a (i, j) indicates (i, j) a element in color component corresponding with information parameter Q to be calculated, that is, works as a (i, j) respectively indicates color component R₁、G₁、B₁、R₂、G₂、B₂In (i, j) a element when, Q respectively indicates corresponding information ginseng Number Q_R、Q_G、Q_B、Q′_R、Q′_G、Q′_B；Wherein, R₁、G₁And B₁For first group of color three-component, R₂、G₂And B₂Divide for second group of color three Amount；Q_R、Q_GAnd Q_BFor the first group information parameter, Q '_R、Q′_GWith Q '_BFor the second group information parameter.

Further, specifically being wrapped using LSS chaos system and Kronecker product generation calculation matrix in the step 3 It includes:

Step 3.1: by the initial value a₀Iteration m in LSS chaos system is brought into parameter r '₁+ Md times, cast out preceding m₁It is a Value obtains the sequence A that length is 1 × Md；Wherein, d is default sampling interval, m₁≥500；

Step 3.2: sequence A is converted to obtain sequence A ' according to formula (15):

A′_i=1-2A_i, i=1,2 ..., Md (15)

Step 3.3: equidistant d sampling is carried out according to formula (16) to sequence A ' progress, obtains sample sequence A ":

A″_k=A '_1+kdK is sampling sequence number, k >=0 (16)

Step 3.4: the chaos sequence S that 4 length are 4 is randomly selected from sequence A "₁、S₂、S₃And S₄, and by chaos sequence Arrange S₁~S₄It is converted into the low-dimensional matrix K that 4 sizes are 2 × 2₁、K₂、K₃And K₄；

Step 3.5: using Kronecker product to low-dimensional matrix K₁And K₂、K₁And K₃、K₂And K₃And K₁And K₄It extends respectively log₂ ^MHigher dimensional matrix K that is secondary, will obtaining₁′、K₂′、K₃' and K₄' it is used as calculation matrix.

Further, optimizing in the step 3 to the calculation matrix specifically includes:

Step 3.6: by calculation matrix K₁' it is converted into calculation matrix φ, the φ ∈ R that size is P × Q^P×Q；Wherein, P=CR × M, Q=N, CR are default compression ratio；

Step 3.7: singular value decomposition, i.e. φ=U Σ V are carried out to calculation matrix φ^T, obtain diagonal matrix Σ；Wherein,Indicate the diagonal matrix of calculation matrix φ, Σ₁=diag (δ₁,δ₂..., δ_r), δ₁≥δ₂≥...≥δ_r> 0, δ₁ ~δ_rIndicate the singular value of calculation matrix φ, r is singular value number；

Step 3.8: all 1's matrix J that construction size is M × N；Calculate the mean value var of diagonal matrix Σ diagonal entry₁, And count Σ₁In >=var₁Singular value number f；The preceding f of order matrix J is arranged multiplied by weighting coefficient t, obtains optimization matrix J₁, In, t > 1；Matrix J will be optimized₁With calculation matrix φ dot product, calculation matrix φ ' after obtaining the first suboptimization；

Step 3.9: singular value decomposition, i.e. φ '=U are carried out again to calculation matrix φ '₁Σ₂V₁ ^T；Wherein,Indicate the diagonal matrix of calculation matrix φ ', Σ₃=diag (δ '₁,δ′₂..., δ '_r),δ′₁~δ '_rIndicate measurement The singular value of matrix φ ', r are singular value number；

Step 3.10: enabling δ '₁=δ '₂=...=δ '_r=1, obtain new diagonal matrix Σ '₂, and according to diagonal matrix Σ′₂Calculation matrix φ after generating final optimization pass₁=U₁Σ′₂V₁；

Step 3.11: referring to step 3.6 to step 3.10, to calculation matrix K₂′、K₃' and K₄' optimize, optimized Calculation matrix φ afterwards₂、φ₃And φ₄。

Further, the step 5 specifically includes:

Step 5.1: by the initial value x₀、y₀、z₀、u₀、v₀And w₀Bring iteration m in sextuple hyperchaotic system into₀+M₁N₁It is secondary, Cast out preceding m₀A value, obtaining 6 sizes is 1 × M₁N₁Sequence X, Y, Z, U, V and W；Wherein, m₀>=500, M₁=M × CR, N₁= N × CR, CR are default compression ratio；

Step 5.2: six chaos sequences X, Y, Z, U, V and W being arranged according to ascending order, obtain six index vector D₁~D₆；

Step 5.3: six chaos sequences X, Y, Z, U, V and W being modified according to formula (19), obtain revised new sequence Arrange XX, YY:

Wherein, XX_(i)、YY_(i)Indicate i-th of element in new chaos sequence XX and YY, X_i、Y_i、Z_i、U_i、V_iAnd W_iRespectively Indicate i-th of element of sequence X, Y, Z, U, V and W, i ∈ [1, M₁N₁]；

Step 5.4: extracting first element value X of new sequence X X, YY respectively_ind1、Y_ind1, and repaired according to formula (20) Just, obtain revised sequence X '_ind1With Y '_ind1:

Step 5.5: if X '_ind1=1, index of reference vector D₁、D₂And D₃Respectively to sequence R₃、G₃And B₃Rearrangement；Such as Fruit X '_ind1=2, index of reference vector D₁、D₂And D₃Respectively to sequence G₃、B₃And R₃Rearrangement；If X '_ind1=3, index of reference to Measure D₁、D₂And D₃To sequence B₃、R₃And G₃Rearrangement；Wherein, R₃、G₃And B₃Sequence is measured for first group of compressed sensing；

Step 5.6: if Y '_ind1=1, index of reference vector D₄、D₅And D₆To sequence R '₃、G′₃With B '₃Rearrangement；If Y′_ind1=2, index of reference vector D₄、D₅And D₆To sequence G '₃、B′₃With R '₃Rearrangement；If Y '_ind1=3, index of reference vector D₄、D₅And D₆To sequence B '₃、R′₃With G '₃Rearrangement；Wherein, R '₃、G′₃With B '₃Sequence is measured for second group of compressed sensing；

Step 5.7: first group of compressed sensing after scramble is measured into sequence R₃、G₃And B₃And second group of compressed sensing is surveyed Measure sequence R '₃、G′₃With B '₃Being separately converted to size is M₁×N₁Matrix R₄、G₄And B₄And R '₄、G′₄With B '₄, wherein R₄、G₄And B₄For first group of Scrambling Matrix, R '₄、G′₄With B '₄For second group of Scrambling Matrix.

Further, the step 6 is specially and is quantified according to formula (21):

Wherein, min and max is the minimum value and maximum value of Scrambling Matrix P respectively, and floor (x), which indicates to calculate, is not more than x Maximum integer；P_(i)With P '_(i)It is i-th of element of Scrambling Matrix P and quantization matrix P ' respectively, that is, works as P_(i)It respectively indicates and sets Random matrix R₄、G₄、B₄、R′₄、G′₄、B′₄In i-th of element when, P '_(i)It is quantization matrix R respectively₅、G₅、B₅、R′₅、G′₅、B′₅ I-th of element；1≤i≤M₁N₁；Wherein, R₅、G₅And B₅For first group of quantization matrix；R′₅、G′₅With B '₅For second group of quantization square Battle array.

Further, the step 11 specifically:

Step 11.1: DWT decomposition being carried out to other two component in third group color three-component, it is equal to respectively obtain size For the Matrix C of (M × N)/4_A、C_H、C_V、C_DWith C '_A、C′_H、C′_V、C′_D；Wherein, other two point in third group color three-component The size of amount is M × N；

Step 11.2: setting CR=0.25 for the compression ratio CR of colored plaintext image；

Step 11.3: by Matrix C_HFour pieces are divided into, 4 matrix L L that size is (M × N)/16 are obtained₁、LH₁、HL₁With HH₁；

Step 11.4: referring to step 11.3, by Matrix C_V、C′_H、C′_VBe divided into four pieces respectively, obtain size be (M × N)/ 16 other 12 matrix L L₂~LL₄、LH₂~LH₄、HL₂~HL₄And HH₂~HH₄；Wherein, LL₁~LL₂、LH₁~LH₂、HL₁~ HL₂、HH₁~HH₂For first group of carrier matrix；LL₃~LL₄、LH₃~LH₄、HL₃~HL₄And HH₃~HH₄For second group of carrier square Battle array.

Further, the step 12 specifically:

Step 12.1: two groups of quantization matrixes being converted according to formula (22) and formula (23) respectively, every group of quantization matrix Obtain transformed two matrixes d and d':

D (i, j)=Q (i, j) mod10 (22)

D ' (i, j)=floor (Q (i, j)/10) (23)

Wherein, Q (i, j) and d (i, j), d ' (i, j) are the matrix obtained after quantization matrix Q and Q to be transformed are converted respectively D, (i, j) a element in d ', i.e., when Q (i, j) is quantization matrix R respectively₅、G₅、B₅、R′₅、G′₅、B′₅In (i, j) it is a When element, d (i, j) is corresponding d respectively₁、d₂、d₃、d₄、d₅、d₆In (i, j) a element, d ' (i, j) is corresponding respectively d₁₁、d₂₂、d₃₃、d₄₄、d₅₅、d₆₆In (i, j) a element；1≤i≤M₁, 1≤j≤N₁；

Step 12.2: according to formula (24) calculating matrix C_AWith C '_APixel average I₁And I₂:

Wherein, C (i, j) is (i, j) a element in Matrix C to be calculated, and I is the pixel average of Matrix C to be calculated, I.e. when C (i, j) is C respectively_AWith C '_AIn (i, j) a element when, I is C respectively_AWith C '_APixel average I₁And I₂；

Step 12.3: extracting Matrix C_AWith C '_AFirst element value E₁And E₂；

Step 12.4: being respectively compared I₁And E₁Size and I₂And E₂Size, and according to comparison result by d₁~d₆With d₁₁~d₆₆It is embedded into first group of carrier matrix and second group of carrier matrix, first group of carrier matrix after being embedded in LL′₁~LL '₂、LH′₁~LH '₂、HL′₁~HL '₂、HH′₁~HH '₂, and second group of carrier matrix LL ' after insertion₃~LL ′₄、LH′₃~LH '₄、HL′₃~HL '₄、HH′₃~HH '₄；

Step 12.5: the matrix L L ' that will be obtained after insertion₁、LH′₁、HL′₁、HH′₁It is reconfigured, obtains new square Battle array C_H1；

Step 12.6: referring to step 12.5, the matrix L L ' that will be obtained after insertion₂~LL '₄、LH′₂~LH '₄、HL′₂~ HL′₄With HH '₂~HH '₄It is reconfigured, obtains new Matrix C_V1、C′_H1With C '_V1；

Step 12.7: by Matrix C_A、C_H1、C_V1、C_DWith C '_A、C′_H1、C′_V1、C′_DSecond is obtained using inverse discrete wavelet transform Ciphertext matrix and third ciphertext matrix.

Further, the step 12.4 specifically:

If E₁≥I₁And E₂≥I₂, d₁It is embedded into LL₁In, d₁₁It is embedded into LH₁In, d₂It is embedded into HL₁In, d₂₂It is embedding Enter to HH₁In, d₃It is embedded into LL₂In, d₃₃It is embedded into LH₂In, d₄It is embedded into LL₃In, d₄₄It is embedded into LH₃In, d₅ It is embedded into HL₃In, d₅₅It is embedded into HH₃In, d₆It is embedded into HL₄In, d₆₆It is embedded into HH₄In；

If E₁≥I₁And E₂<I₂, d₁It is embedded into LL₂In, d₁₁It is embedded into LH₂In, d₂It is embedded into LL₃In, d₂₂It is embedding Enter to LH₃In, d₃It is embedded into HL₃In, d₃₃It is embedded into HH₃In, d₄It is embedded into HL₄In, d₄₄It is embedded into HH₄In, d₅ It is embedded into LL₁In, d₅₅It is embedded into LH₁In, d₆It is embedded into HL₁In, d₆₆It is embedded into HH₁In；

If E₁<I₁And E₂≥I₂, d₁It is embedded into LL₃In, d₁₁It is embedded into LH₃In, d₂It is embedded into HL₃In, d₂₂It is embedding Enter to HH₃In, d₃It is embedded into HL₄In, d₃₃It is embedded into HH₄In, d₄It is embedded into LL₁In, d₄₄It is embedded into LH₁In, d₅ It is embedded into HL₁In, d₅₅It is embedded into HH₁In, d₆It is embedded into LL₂In, d₆₆It is embedded into LH₂In；

If E₁<I₁And E₂<I₂, d₁It is embedded into HL₄In, d₁₁It is embedded into HH₄In, d₂It is embedded into LL₁In, d₂₂Insertion To LH₁In, d₃It is embedded into HL₁In, d₃₃It is embedded into HH₁In, d₄It is embedded into LL₂In, d₄₄It is embedded into LH₂In, d₅It is embedding Enter to LL₃In, d₅₅It is embedded into LH₃In, d₆It is embedded into HL₃In, d₆₆It is embedded into HH₃In.

Beneficial effects of the present invention:

The two width color image compression encryption methods provided by the invention based on two dimensional compaction perception have below beneficial to effect Fruit:

(1) by the way that two width colour plaintext images are embedded into the identical color host image of an auxiliary scale cun, using two dimension Compressed sensing carries out compression encryption, scramble, quantization from color three-component of two different directions to colored plaintext image, and will Encrypted color image is embedded into color host image, finally obtains visual security of the size as plaintext image Ciphertext image realizes picture appearance safety and data safety, and the data volume of encrypted image is reduced using two dimensional compaction perception, makes Obtained ciphertext size is smaller, reduces the runing time of the algorithm, and saves transmission bandwidth and memory space.

(2) close by the way that the two group information parameters and cryptographic Hash closely related with plaintext are embedded into color host image Key parameter is transmitted directly to recipient with the ciphertext image of visual security, in this way when encrypting simultaneously to high-volume picture, keeps away Exempt to save transmission bandwidth and memory space to the additional transmissions and storage of relevant key parameter in plain text, be particularly suited for Resource constrained environment, and improve the safety of Encryption Algorithm of the present invention.

(3) by carrying out shuffle operation to the measured value of plaintext image during compressing and encrypting, adjacent member is reduced The correlation of element.The index vector as used in scramble is to be generated by chaos sequence, and the generation of chaos sequence and plaintext are tight Close correlation is that random, different plaintext image has different scrambles as a result, can effectively support this guarantees scrambling process Imperial known-plaintext and chosen -plain attact.Moreover, telescopiny is related with color host image, the letter of color host image is utilized Cease the insertion random to ciphertext.Therefore the random operation in scramble and insertion further improves the safety of algorithm.

Detailed description of the invention

Fig. 1 is the stream of the two width color image compression encryption methods provided in an embodiment of the present invention based on two dimensional compaction perception One of journey schematic diagram；

Fig. 2 is the stream of the two width color image compression encryption methods provided in an embodiment of the present invention based on two dimensional compaction perception The two of journey schematic diagram；

Fig. 3 is the process of the initial value of generation LSS chaos system provided in an embodiment of the present invention and sextuple hyperchaotic system Schematic diagram；

Fig. 4 is the two width color images based on two dimensional compaction perception that inventive embodiments provided in an embodiment of the present invention provide The three of the flow diagram of compression encryption method；

Fig. 5 is the encryption of colored plaintext image Lena and Baboon provided in an embodiment of the present invention and the signal of decrypted result Figure；

Fig. 6 is the key sensitivity tests result schematic diagram in ciphering process provided in an embodiment of the present invention；

Fig. 7 is the key sensitivity tests result schematic diagram in decrypting process provided in an embodiment of the present invention；

Fig. 8 is influence schematic diagram of the salt-pepper noise of varying strength provided in an embodiment of the present invention to decrypted result；

Fig. 9 is the decrypted image schematic diagram in the case of no shearing provided in an embodiment of the present invention；

Figure 10 is the decrypted image schematic diagram in the case of shearing 1/8th provided in an embodiment of the present invention.

Specific embodiment

To make the object, technical solutions and advantages of the present invention clearer, below in conjunction with attached in the embodiment of the present invention Figure, technical solution in the embodiment of the present invention are explicitly described, it is clear that described embodiment is a part of the invention Embodiment, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art are not making wound Every other embodiment obtained under the premise of the property made labour, shall fall within the protection scope of the present invention.

As shown in Figure 1, the two width color image compression encryption sides provided in an embodiment of the present invention based on two dimensional compaction perception Method, comprising the following steps:

S11: the first width colour plaintext image P is extracted respectively₁With the second width colour plaintext image P₂Color three-component, note For first group of color three-component and second group of color three-component, and two groups of colors three are divided respectively using two-dimensional discrete wavelet conversion Amount carries out rarefaction；Wherein, P₁And P₂Size be M × N and M=N；

Specifically, in the embodiment of the present invention, by the first width colour plaintext image P₁Color three-component be denoted as first group of face Color three-component: R₁、G₁And B₁；By the second width colour plaintext image P₂Color three-component be denoted as second group of color three-component: R₂、G₂ And B₂.The color three-component refers to the Red component, Green component and Blue component of color image.

S12: it calculates separately to obtain according to two groups of color three-components of two width colour plaintext images bright with the two width colour Texts and pictures are as relevant first group information parameter and the second group information parameter；

Specifically, in the embodiment of the present invention, the information parameter being calculated according to first group of color three-component is denoted as One group information parameter；The information parameter being calculated according to second group of color three-component is denoted as the second group information parameter.Due to The first group information parameter and the second group information are obtained according to the color three-component of two width color images, therefore two groups of letters It ceases parameter and plaintext image is closely related, be embedded into color host image for subsequent a part as key information.

S13: calculation matrix is generated using LSS chaos system and Kronecker product, and the calculation matrix is optimized；

Specifically, common one-dimensional chaos system has Logistic mapping (formula (1)) and Sine mapping (formula (2)):

x′_n+1=F_L(u,x′_n)=u × x_n×(1-x′_n) (1)

x″_n+1=F_S(r,x″_n)=r × sin (π × x "_n)/4 (2)

Wherein, when u ∈ (3.89,4] and when (0,1) r ∈, Logistic mapping and Sine mapping have chaotic behavior.

Simple in view of above-mentioned traditional One Dimensional Chaotic Maps structure, randomness is limited, and the data distribution of output is uneven, no Few experts and scholars propose some improved, with more advantageous characteristic systems.Logistic mapping is mutually tied with Sine mapping It closes, obtain a new accountant rule: Logistic-Sine maps (LSS), as shown in formula (3):

Wherein r ' ∈ (0,4] be control parameter, be key a part.Selection uses formula (3) institute in the embodiment of the present invention The new accountant rule shown, available bigger chaos range, and the sequence exported is evenly distributed in [0,1], than tradition 1 dimensional Logistic Map or Sine mapping chaos performance it is more preferable.

Kronecker product (Kronecker Product, KP) is the operation between a kind of matrix of two arbitrary sizes, it is The special shape of tensor product, with the naming of Germany mathematics person's Leopold Kronecker.Two sizes are respectively m × n Formula (4) are defined as with the matrix A of p × q and the KP of B:

Wherein, matrixSize be mp × nq, and meet

Such as matrixThenResult are as follows:

Such asThenResult are as follows:

KP has calculating speed fast, the low feature of computation complexity.If A and B are nonsingular, and under having (on) Triangle, band-like, symmetrical, positive definite, random, Teoplitz, displacement and the properties such as orthogonal, thenAnd it is nonsingular, Under having (on) triangle, band-like, symmetrical, positive definite, random, Teoplitz, displacement and the properties such as orthogonal.By above-mentioned property, Low-dimensional is random, linear independence or orthogonal base vectors can expand to random higher-dimension, linear independence or orthogonal matrix by KP.At this In inventive embodiments, chaotic maps are applied to generate low-dimensional Seed Matrix, then utilize KP by low-dimensional matrix-expand to higher-dimension Matrix generates the calculation matrix of compressed sensing (CS), and something which increases the execution efficiencys of the algorithm.

S14: two dimensional compaction sense is carried out to two groups of color three-components after rarefaction respectively using the calculation matrix after optimization Know, obtains corresponding first group of compressed sensing measurement sequence and second group of compressed sensing measurement sequence；

Specifically, in the embodiment of the present invention, calculation matrix is denoted as the three-component compressed sensing result of first group of color First group of compressed sensing result；Calculation matrix is denoted as second group of compression sense to the three-component compressed sensing result of second group of color Know result.Then, the sequence that first group of compressed sensing result converts is denoted as first group of compressed sensing measurement sequence；By The sequence that two groups of compressed sensing results convert is denoted as second group of compressed sensing measurement sequence.

S15: shuffle operation is carried out to two groups of compressed sensing measurement sequences respectively using sextuple hyperchaotic system, is corresponded to First group of Scrambling Matrix and second group of Scrambling Matrix, and by element value in two groups of Scrambling Matrixes be less than preset threshold TS element Value is changed to 0, obtains revised two groups of Scrambling Matrixes；

Specifically, sextuple hyperchaotic system has complicated phase space and better dynamic characteristic and randomness, and gathers around There is more complicated dynamic behavior, which has the structure more more complicated than one-dimensional chaos system, so the embodiment of the present invention Sextuple hyperchaotic system is selected to can be improved the safety of encryption method.Shown in the model of sextuple hyperchaotic system such as formula (5):

Wherein, x, y, z, u, v, w represent the state variable of sextuple hyperchaotic system；A, b, c, d, e, f, g are system parameter； Work as a=10, b=4, c=20, d=-2.5, e=4, when f=5, g=20, system is in hyperchaos state.

S16: revised two groups of Scrambling Matrixes are quantified respectively, element value is transformed between 0 to 255, is obtained Corresponding first group of quantization matrix and second group of quantization matrix；

S17: the color three-component of the color host image F ' after extracting image normalization is denoted as third group color three and divides Amount；Wherein, the size of color host image F ' is M × N；

Specifically, the image normalization in the embodiment of the present invention specifically:

The pixel value of color host image F is normalized between [10,245] according to formula (6), the coloured silk after being normalized Color carrier image F ':

Wherein, parameter A=10, μ=245；Expression rounds up to the value of x；Such as:It is colored The size of carrier image F is M × N.

Carrying out image normalization to color host image in this step is in order to prevent after embedding data, and data are overflowed, and are led Cause obtained visual security image visually variant.

S18: 256 cryptographic Hash are generated using the SHA256 function of the first width colour plaintext image, and described 256 are breathed out Uncommon every 8 one group of value is converted into 32 decimal number k₁, k₂..., k₃₂, 32 decimal numbers are denoted as cryptographic Hash K；

S19: will the first group information parameter relevant to two width colour plaintext images and the second group information parameter according to setting Parameter Switch rule be respectively converted into first group of integer set and second group of integer set, wherein first group of set of integers It closes and second group of integer set includes L integer；

S110: (32+2L) a pixel value before one of component in third group color three-component is replaced with described Cryptographic Hash K and 2L integer obtain the first ciphertext matrix containing key parameter information；

S111: other two component in third group color three-component is subjected to DWT decomposition, and the square that will be obtained after decomposition Battle array carries out piecemeal, obtains corresponding first group of carrier matrix and second group of carrier matrix；

Specifically, DWT refers to wavelet transform.

S112: first group of quantization matrix and second group of quantization matrix are respectively embedded in two groups of carrier matrixs, obtained To the second ciphertext matrix and third ciphertext matrix containing ciphertext image information；

S113: first ciphertext matrix, the second ciphertext matrix and third ciphertext matrix are combined, and are obtained and two width The corresponding visual security image of colored plaintext image.

It should be noted that each step in the embodiment of the present invention is in the absence of conflict, part steps can be simultaneously It executes, can also successively execute.

The two width color image compression encryption methods provided in an embodiment of the present invention based on two dimensional compaction perception, compared to existing Some image encryption methods, have the advantage that

(3) by carrying out shuffle operation to the measured value of plaintext image during compressing and encrypting, adjacent member is reduced The correlation of element.

On the basis of the above embodiments, in conjunction with Fig. 2 to Fig. 4, the embodiment of the present invention provides another and is based on two dimensional compaction Two width color image compression encryption methods of perception, comprising the following steps:

S21: P is extracted respectively₁And P₂Color three-component, and using two-dimensional discrete wavelet conversion respectively to two groups of colors three Component carries out rarefaction；

Specifically, specific to two groups of color three-components progress rarefactions respectively using two-dimensional discrete wavelet conversion (2D DWT) Are as follows: according to formula (7) to R₁、G₁、B₁And R₂、G₂、B₂Carry out rarefaction:

Wherein, ψ is sparse basis, by R₁、G₁、B₁Matrix after sparse is denoted as R '₁、G′₁、B′₁；By R₂、G₂、B₂After sparse Matrix is denoted as R '₂、G′₂、B′₂.Wherein, R '₁、G′₁、B′₁、R′₂、G′₂、B′₂Size is M × N.

S22: it is calculated and the first group information parameter and the second group information parameter according to two groups of color three-components；

Specifically, the first group information parameter and the second group information parameter are calculated according to formula (8):

Wherein, a (i, j) indicates (i, j) a element in color component corresponding with information parameter Q to be calculated, that is, works as a (i, j) respectively indicates color component R₁、G₁、B₁、R₂、G₂、B₂In (i, j) a element when, Q respectively indicates corresponding information ginseng Number Q_R、Q_G、Q_B、Q′_R、Q′_G、Q′_B；Wherein, Q_R、Q_GAnd Q_BFor the first group information parameter, Q '_R、Q′_GWith Q '_BFor the second group information ginseng Number.

To calculate information parameter Q_RFor: parameter Q is calculated according to formula (8)_R:

Wherein, a (i, j) indicates R₁(i, j) a element.

S23: P is utilized₁SHA256 function generate 256 cryptographic Hash, and will be every 8 one group of 256 cryptographic Hash turn Turn to 32 decimal number k₁, k₂..., k₃₂, 32 decimal numbers are denoted as cryptographic Hash K；

S24: the initial value of sextuple hyperchaotic system and LSS chaos system is generated；

Specifically, as shown in figure 3, the initial value for generating sextuple hyperchaotic system specifically includes:

As shown in figure 3, the initial value for generating LSS chaos system specifically includes:

Wherein, (t₆,t₇) ∈ (0,1) be preset-key parameter.

S25: calculation matrix is generated using LSS chaos system and Kronecker product, and the calculation matrix is optimized；

Specifically, calculation matrix is generated using LSS chaos system and Kronecker product to specifically include:

Step C1: by the initial value a₀Iteration m in LSS chaos system is brought into parameter r '₁+ Md times, cast out preceding m₁It is a Value obtains the sequence A that length is 1 × Md；Wherein, d is default sampling interval, m₁≥500；

Specifically, when sampling interval d is sufficiently large, the chaos sequence of generation is chaos that is approximately independent, sampling The matrix of sequence construct meets RIP (Restricted Isometry Property in very maximum probability；Limited equidistant property) Condition.LSS chaos system has been introduced in the above-described embodiments, and details are not described herein again.

Step C2: sequence A is converted to obtain sequence A ' according to formula (15):

A′_i=1-2A_i, i=1,2 ..., Md (15)

Specifically, since new sequence A ' has zero-mean and zero symmetrically, this step can be such that subsequent step constructs The calculation matrix more random distribution come.

Step C3: equidistant d sampling is carried out according to formula (16) to sequence A ' progress, obtains sample sequence A ":

A″_k=A '_1+kdK is sampling sequence number, k >=0 (16)

Step C4: the chaos sequence S that 4 length are 4 is randomly selected from sequence A "₁、S₂、S₃And S₄, and by chaos sequence S₁~S₄It is converted into the low-dimensional matrix K that 4 sizes are 2 × 2₁、K₂、K₃And K₄；

Step C5: using Kronecker product to low-dimensional matrix K₁And K₂、K₁And K₃、K₂And K₃And K₁And K₄It extends respectively log₂ ^MHigher dimensional matrix K that is secondary, will obtaining₁′、K₂′、K₃' and K₄' it is used as calculation matrix.

Sparse signal is projected lower dimensional space from higher dimensional space by calculation matrix, in the condition for guaranteeing signal reconstruction precision Under, it is desirable that the lower projector space dimension the better.

The nonzero coefficient of sparse signal is concentrated mainly on low-frequency band, and zero coefficient or is concentrated mainly on height close to zero coefficient Frequency band.Therefore, it by increasing the measurement coefficient of the first half of calculation matrix, can be obtained within the identical sampling time more Information, to accurately rebuild original signal.However, increasing the coefficient of calculation matrix first half will inevitably reduce The incoherence of matrix.According to the property of Singular Value Decomposition Using, maximum singular value is smaller, and the incoherence of matrix is better.It is real It tests the result shows that improved singular value calculation matrix has RIP property more better than original matrix, and rebuilds effect and obviously mention It is high.The characteristics such as singular value is with good stability, constant rate, rotational invariance, therefore calculation matrix is carried out unusual Value decomposes the purpose, it can be achieved that calculation matrix optimization.Singular Value Decomposition Using method is as follows:

Assuming that matrix A ∈ R^m×n, there are orthogonal matrix U ∈ R^m×n、V∈R^m×nMake with diagonal matrix Σ:

A=U Σ V^T (17)

Wherein,And Σ₁=diag (δ₁, δ₂..., δ_r), cornerwise element δ in sequence₁> > δ₂ > > ... δ_r> 0, r=rank (A) arrangement.Here it is matrix singular value decomposition theorems, wherein δ₁, δ₂..., δ_rFor matrix A Singular value, formula (17) is known as the singular value decomposition of matrix A.

In the embodiment of the present invention, the calculation matrix is optimized and is specifically included:

Step C6: by calculation matrix K₁' it is converted into calculation matrix φ, the φ ∈ R that size is P × Q^P×Q；Wherein, P=CR × M, Q=N, CR are default compression ratio；

Step C7: singular value decomposition, i.e. φ=U Σ V are carried out to calculation matrix φ^T, obtain diagonal matrix Σ；Wherein,Indicate the diagonal matrix of calculation matrix φ, Σ₁=diag (δ₁,δ₂..., δ_r), δ₁≥δ₂≥…≥δ_r> 0, δ₁~ δ_rIndicate the singular value of calculation matrix φ, r is singular value number；

Step C8: all 1's matrix J that construction size is M × N；Calculate the mean value var of diagonal matrix Σ diagonal entry₁, and Count Σ₁In >=var₁Singular value number f；The preceding f of order matrix J is arranged multiplied by weighting coefficient t (t > 1), obtains optimization matrix J₁； Matrix J will be optimized₁With calculation matrix φ dot product, calculation matrix φ ' ' after obtaining the first suboptimization；

Step C9: singular value decomposition, i.e. φ '=U are carried out again to calculation matrix φ '₁Σ₂V₁ ^T；Wherein, Indicate the diagonal matrix of calculation matrix φ ', Σ₃=diag (δ '₁,δ′₂..., δ '_r),δ′₁~δ '_rIndicate calculation matrix φ's ' Singular value, r are singular value number；

Step C10: δ ' is enabled₁=δ '₂=...=δ '_r=1, obtain new diagonal matrix Σ '₂, and according to diagonal matrix Σ '₂ Calculation matrix φ after generating final optimization pass₁=U₁Σ′₂V₁；

Step C11: referring to step C6 to step C10, to calculation matrix K₂′、K₃' and K₄' optimize, after obtaining optimization Calculation matrix φ₂、φ₃And φ₄。

Specifically, calculation matrix φ₁、φ₂、φ₃And φ₄Size is P × Q.

S26: two dimensional compaction sense is carried out to two groups of color three-components after rarefaction respectively using the calculation matrix after optimization Know, obtains corresponding first group of compressed sensing measurement sequence and second group of compressed sensing measurement sequence；

Specifically, two dimensional compaction is carried out to two groups of color three-components after rarefaction respectively using the calculation matrix after optimization It perceives (2D CS) specifically:

To R '₁、G′₁、B′₁With R '₂, G '₂, B '₂Two dimensional compaction perception is carried out according to formula (18):

Wherein,WithIt is first group of compressed sensing result and second group of pressure respectively Contracting sensing results；φ₁、φ₂、φ₃And φ₄For the calculation matrix after optimization.

Then, by matrixWithIt is converted into sequence R₃、G₃、B₃With R '₃、G′₃、B′₃.Its In, R₃、G₃、B₃Sequence is measured for first group of compressed sensing；R′₃、G′₃、B′₃Sequence is measured for second group of compressed sensing.R₃、G₃、 B₃、R′₃、G′₃、B′₃Length be M₁×N₁, wherein M₁=M × CR, N₁=N × CR；CR is default compression ratio.

S27: shuffle operation is carried out to two groups of compressed sensing measurement sequences respectively using sextuple hyperchaotic system, and to two groups Scrambling Matrix is modified；

Specifically, sextuple hyperchaotic system has been introduced in the above-described embodiments, and details are not described herein again.This step specifically:

Step D1: by the initial value x₀、y₀、z₀、u₀、v₀And w₀Bring iteration m in sextuple hyperchaotic system into₀+M₁N₁It is secondary, Cast out preceding m₀A value, obtaining 6 sizes is 1 × M₁N₁Sequence X, Y, Z, U, V and W；Wherein, m₀>=500, M₁=M × CR, N₁= N × CR, CR are default compression ratio；

Specifically, sequence X, Y, Z, U, V and W can be respectively indicated are as follows: With

Step D2: six chaos sequences X, Y, Z, U, V and W are arranged according to ascending order, obtain six index vector D₁~D₆；

Step D3: six chaos sequences X, Y, Z, U, V and W are modified according to formula (19), obtain revised new sequence Arrange XX, YY:

Step D4: first element value X of new sequence X X, YY is extracted respectively_ind1、Y_ind1, and repaired according to formula (20) Just, obtain revised sequence X '_ind1With Y '_ind1:

Step D5: if X '_ind1=1, index of reference vector D₁、D₂And D₃Respectively to sequence R₃、G₃And B₃Rearrangement；If X′_ind1=2, index of reference vector D₁、D₂And D₃Respectively to sequence G₃、B₃And R₃Rearrangement；If X '_ind1=3, index of reference vector D₁、D₂And D₃To sequence B₃、R₃And G₃Rearrangement；Wherein, R₃、G₃And B₃Sequence is measured for first group of compressed sensing；

Step D6: if Y '_ind1=1, index of reference vector D₄、D₅And D₆To sequence R '₃、G′₃With B '₃Rearrangement；If Y′_ind1=2, index of reference vector D₄、D₅And D₆To sequence G '₃、B′₃With R '₃Rearrangement；If Y '_ind1=3, index of reference vector D₄、D₅And D₆To sequence B '₃、R′₃With G '₃Rearrangement；Wherein, R '₃、G′₃With B '₃Sequence is measured for second group of compressed sensing；

Step D7: first group of compressed sensing after scramble is measured into sequence R₃、G₃And B₃And second group of compressed sensing is surveyed Measure sequence R '₃、G′₃With B '₃Being separately converted to size is M₁×N₁Matrix R₄、G₄And B₄And R '₄、G′₄With B '₄, wherein R₄、G₄And B₄For first group of Scrambling Matrix, R '₄、G′₄With B '₄For second group of Scrambling Matrix；

Specifically, first group of compressed sensing after scramble is measured into sequence R₃、G₃And B₃Matrix after conversion is denoted as first group Scrambling Matrix R₄、G₄And B₄；Second group of compressed sensing after scramble is measured into sequence R '₃、G′₃With B '₃Matrix after conversion is denoted as Second group of Scrambling Matrix R '₄、G′₄With B '₄。

Step D8: the element value by element value in two groups of Scrambling Matrixes less than preset threshold TS is changed to 0, obtains revised Two groups of Scrambling Matrixes.

S28: revised two groups of Scrambling Matrixes are quantified respectively；

Specifically, quantified according to formula (21):

With Scrambling Matrix R₄Quantified to obtain quantization matrix R₅For:

Wherein, min is Scrambling Matrix R₄Minimum value, max is Scrambling Matrix R₄Maximum value, R_4(i)And R_5(i)It is respectively Scrambling Matrix R₄With quantization matrix R₅I-th of element.

S29: the color three-component of the color host image F ' after extracting image normalization is denoted as third group color three and divides Amount；Wherein, the size of color host image F ' is M × N；

Specifically, in the embodiment of the present invention, third group color three-component is expressed as F '_R、F′_GWith F '_B。

S210: the first group information parameter and the second group information parameter are respectively converted into according to the Parameter Switch rule of setting First group of integer set and second group of integer set, wherein first group of integer set and second group of integer set are equal Include L integer；

Specifically, as an embodiment, with the Q in the first group information parameter_RIt is embedded into color host image F′_RFor in component: for example, parameter Q_RIt is 0.589486987492596, then decimal system parameter Q_RIt is divided into 8 integers, i.e., Q_R=(58,94,86,98,74,92,59,6), with same method the Q being calculated_R、Q_G、Q_B、Q′_R、Q′_G、Q′_BIt resolves into 8 integers then obtain 48 integers, use z₁-z₄₈It indicates.Then 32 decimal system cryptographic Hash K z₄₉-z₈₀It indicates.Finally 80 decimal integers are obtained.I.e. in present embodiment, L=24.

S211: (32+2L) a pixel value before one of component in third group color three-component is replaced with described Cryptographic Hash K and 2L integer obtain the first ciphertext matrix containing key parameter information；

Specifically, with the Q in the first group information parameter_RIt is embedded into the F ' of color host image_RFor in component: extracting color Color carrier image F '_RPreceding 80 pixel values of component, with z '₁-z′₈₀It indicates.By z obtained in step S210₁-z₈₀It is substituted into coloured silk Color carrier image F '_RPreceding 80 pixel value z ' of component₁-z′₈₀In to get arrive the first ciphertext containing important parameter key information Matrix F "_R, size is M × N, and key insertion terminates.

S212: other two component in third group color three-component is subjected to DWT decomposition, and the square that will be obtained after decomposition Battle array carries out piecemeal, obtains corresponding first group of carrier matrix and second group of carrier matrix；

Specifically, as an embodiment, other two component in third group color three-component is set as F '_G、 F′_B, this step specifically:

Step E1: to color component F '_G、F′_BDWT decomposition is carried out, the Matrix C that size is (M × N)/4 is respectively obtained_A、 C_H、C_V、C_DWith C '_A、C′_H、C′_V、C′_D；Wherein, F '_G、F′_BSize is M × N；

Step E2: CR=0.25 is set by the compression ratio CR of colored plaintext image；

Specifically, quantization matrix R at this time₅、G₅、B₅With R '₅、G′₅、B′₅Size be (M × N)/16.This step institute To set CR=0.25 for compression ratio CR, in order to which the ciphertext image for obtaining compression measurement is ideally embedded into colored carriers figure As in F.

Step E3: by Matrix C_HFour pieces are divided into, 4 matrix L L that size is (M × N)/16 are obtained₁、LH₁、HL₁And HH₁；

Step E4: referring to step E3, by Matrix C_V、C′_H、C′_VIt is divided into four pieces respectively, obtaining size is (M × N)/16 Other 12 matrix L L₂~LL₄、LH₂~LH₄、HL₂~HL₄And HH₂~HH₄；Wherein, LL₁~LL₂、LH₁~LH₂、HL₁~HL₂、 HH₁~HH₂For first group of carrier matrix；LL₃~LL₄、LH₃~LH₄、HL₃~HL₄And HH₃~HH₄For second group of carrier matrix.

S213: first group of quantization matrix and second group of quantization matrix are respectively embedded in two groups of carrier matrixs, obtained To the second ciphertext matrix and third ciphertext matrix containing ciphertext image information；

Specifically, this step specifically:

Step F1: two groups of quantization matrixes are converted according to formula (22) and formula (23) respectively, every group of quantization matrix is equal Obtain transformed two matrixes d and d':

D (i, j)=Q (i, j) mod10 (22)

D ' (i, j)=floor (Q (i, j)/10) (23)

With quantization matrix R₅For: it is converted according to equation (22) and (23), obtains two matrix d₁And d₁₁。

d₁(i, j)=R₅(i,j)mod10

d₁₁(i, j)=floor (R₅(i,j)/10)

Wherein, d₁(i,j)、d₁₁(i, j) and R₅(i, j) is matrix d respectively₁、d₁₁And R₅In the position (i, j) element, 1≤i ≤M₁, 1≤j≤N₁。

Step F2: according to formula (24) calculating matrix C_AWith C '_APixel average I₁And I₂:

With Matrix C_AFor: it is calculated according to formula (24):

Wherein, C_A(i, j) is Matrix C_A(i, j) a element；I₁It is C_APixel average.

Step F3: Matrix C is extracted_AWith C '_AFirst element value E₁And E₂；

Step F4: it is respectively compared I₁And E₁Size and I₂And E₂Size, and according to comparison result by d₁~d₆And d₁₁ ~d₆₆It is embedded into first group of carrier matrix and second group of carrier matrix；

Specifically, as an embodiment, if E₁≥I₁And E₂≥I₂, d₁It is embedded into LL₁In, d₁₁It is embedded into LH₁In, d₂It is embedded into HL₁In, d₂₂It is embedded into HH₁In, d₃It is embedded into LL₂In, d₃₃It is embedded into LH₂In, d₄Insertion To LL₃In, d₄₄It is embedded into LH₃In, d₅It is embedded into HL₃In, d₅₅It is embedded into HH₃In, d₆It is embedded into HL₄In, d₆₆It is embedding Enter to HH₄In；

Finally, the matrix for obtaining insertion cipher-text information is denoted as: LL '₁-LL′₄、LH′₁-LH′₄、HL′₁-HL′₄With HH '₁-HH ′₄, size is (M × N)/16.Wherein, LL '₁~LL '₂、LH′₁~LH '₂、HL′₁~HL '₂、HH′₁~HH '₂After insertion First group of carrier matrix；LL′₃~LL '₄、LH′₃~LH '₄、HL′₃~HL '₄、HH′₃~HH '₄For second group of carrier square after insertion Battle array.

Step F5: the matrix L L ' that will be obtained after insertion₁、LH′₁、HL′₁、HH′₁It is reconfigured, obtains new matrix C_H1；

Step F6: referring to step F5, the matrix L L ' that will be obtained after insertion₂~LL '₄、LH′₂~LH '₄、HL′₂~HL '₄With HH′₂~HH '₄It is reconfigured, obtains new Matrix C_V1、C′_H1With C '_V1；

Specifically, by matrix L L '₂、LH′₂、HL′₂、HH′₂It is reconfigured, obtains new Matrix C_V1；By matrix L L ′₃、LH′₃、HL′₃、HH′₃Reconfigured, obtain new Matrix C '_H1；By matrix L L '₄、LH′₄、HL′₄、HH′₄It carries out again Combination, obtain new Matrix C '_V1.Wherein, C_H1、C_V1、C′_H1With C '_V1Size is (M × N)/4.

Step F7: by Matrix C_A、C_H1、C_V1、C_DWith C '_A、C′_H1、C′_V1、C′_DMatrix is obtained using inverse discrete wavelet transform F″_GWith F "_B；Wherein, F "_GFor the second ciphertext matrix；F″_BFor third ciphertext matrix.

Specifically, F "_GWith F "_BSize is M × N.

S214: first ciphertext matrix, the second ciphertext matrix and third ciphertext matrix are combined, and are obtained and two width The corresponding visual security image of colored plaintext image.

Specifically, by F "_R、F″_G、F″_BThree-component is combined, and the visual security ciphertext graph containing cipher-text information is obtained As C.Ciphering process terminates.

In order to verify the effective of the two width color image compression encryption methods provided by the invention based on two dimensional compaction perception Property, the present invention provides following experiment.

Experiment porch and experiment parameter: mentioned algorithm is verified using Matlab R2016a in 2.5GHz CPU and 4GB The encryption and decryption effect on PC deposited, operating system is Microsoft Windows 10.Size is 512 × 512 Color Lena image and colour Baboon image be used as plaintext image, color host image for identical size Lake.It uses Parameter setting it is as follows: t₁=0.35, t₂=0.15, t₃=0.25, t₄=0.45, t₅=0.56, t₆=0.51, t₇=0.45； Threshold value TS=30, sampling interval d are 20, compression ratio CR=0.25.In the DWT of color host image is decomposed, use " Haar " Wavelet function.In the reconstruction process of CS, using ONSL₀Method.Wherein, encryption and decrypted result are illustrated in fig. 5 shown below.

Fig. 5 is the encryption of colored plaintext image Lena and Baboon, decrypted result.In Fig. 5, (a) and (b) is colored respectively Plaintext image Lena and Baboon, are (c) color host image Lake, are (d) colored plaintext image Lena and Baboon respectively Red, Green, Blue three-component compression measurement after ciphertext image, (e) be corresponding visual security ciphertext image, (f) It (g) is its corresponding decrypted image.

As can be seen from Figure 5: (1): (d) is the ciphertext graph after the compression measurement of colored plaintext image Lena and Baboon Picture, they are the images of similar noise, when they are when transmission over networks stores, it is easy to it is found and is attacked by hacker, it cannot Effective protection image data safety；(2): (e) is the ciphertext image of corresponding visual security, they are meaningful and from appearance On be not to be regarded as ciphertext image, effectively hide cipher-text information, protect data safety, have relatively high appearance safety level It not, is 33.35dB with the PSNR value of figure (c), visual effect is good；(3): (f) and (g) shown in decrypted image and (a) and (b) plaintext image shown in is visually almost the same, is visually difficult to find out that the difference of the two, corresponding PSNR value are respectively 33.56dB and 30.01dB, it is therefore proposed that Encryption Algorithm have it is good encryption and decryption effect, image data can be obtained Safety and appearance safety, and the visual security password image for encrypting generation is identical as original image size, does not need additional Transmission bandwidth and memory space.

One, key space is analyzed

It is well known that the size of key space reflects the difficulty and complexity of successful attack cryptographic system, therefore effectively Image Encryption Scheme need sufficiently large key space to resist brute force attack.The key of Encryption Algorithm proposed by the invention Specifically include that 256 cryptographic Hash K that (1) is obtained from 256 hash function of SHA of plaintext image；(2) given parameter t₁-t₇； (3) key Q relevant to plaintext image information_R、Q_G、Q_B、Q’_R、Q’_G、Q’_B.In addition, the DWT of threshold value TS and at least 37 seed types Existing wavelet filter can also be used as key.

If the computational accuracy of computer is 10^-14, then key space is about 37²×(10¹⁴)⁷=37²×10⁹⁸>2²⁹⁴Such as Fruit adds 256 cryptographic Hash K, and whole key is considerably beyond 2¹⁰⁰, this shows that the key space of this paper is sufficiently large, Neng Gouyou Effect resists exhaustive attack, therefore safety with higher.

Two, key sensitivity analysis

Key sensibility refers to that the ciphertext image obtained after slight change and former ciphertext image occurs when primary key Between difference, difference is bigger, illustrates that key sensibility is higher.Equally, it is obtained with the key decryption ciphertext image after slight change The decrypted image arrived and plaintext image difference are bigger, and decruption key sensibility is also better.The Encryption Algorithm of proposition should be to key Variation it is very sensitive.This experiment is used as the color Lena image that size is 512 × 512 with colour Baboon image to be schemed in plain text Picture, (a), (b) are shown in Fig. 5, and the colored Lake image that size is 512 × 512 is as carrier image, (c) institute in Fig. 5 Show, this experiment will be tested in terms of the two from encryption key sensibility and decruption key sensibility.

1) encryption key sensibility

Firstly, this experiment tests the sensibility of encryption key.The key of selection are as follows: t₁, t₂With 256 cryptographic Hash Key, Change a key parameter every time, other parameters are constant.In ciphering process, pass through addition 10^-14To change t₁, t₂, after modification Key be T₁=t₁+10^-14, T₂=t₂+10^-14, a certain position of 256 cryptographic Hash K is then modified, a new Key is obtained₁, Wherein, Key and Key₁It is as follows:

Key=[d3d923762687e3694e7438d72afb166936f282c8f3881c5271d42550 588dfa 34]

Key₁=[d3d923762687e3694e7438d72afb166936f282c8f3881c5271d42550 588df a39]

Encrypted image is as shown in Figure 6.Fig. 6 is the key sensitivity tests result in ciphering process: where (a) be using The visual security image obtained under correct key；(b)~(d) is using cipher key T respectively₁、T₂、Key₁Obtained visual security figure Picture.From fig. 6, it can be seen that when key is changed, from the appearance from the point of view of, corresponding encrypted image and color host image are regarding Without apparent difference in feel, this is because color host image plays an important role in the ciphertext image of visual security, therefore The modification key pair encryption stage has little effect.

2) decruption key sensibility

Ciphertext image (Fig. 6 (a)) is decrypted in this experiment using above-mentioned modified key, to test in decrypting process Key sensibility.As a result as shown in Figure 7:

Fig. 7 is the key sensitivity tests result in decrypting process: where (a) is using the decryption figure under correct code key As Lena and Baboon；(b)~(d) is using cipher key T respectively₁、T₂、Key₁Obtained decrypted image.As can be seen from Figure 7: (1): plaintext image can be come out with correct key recovery, visually be difficult to find out decrypted image and the in plain text difference of the two. (2): when key has atomic small variation, obtained decrypted image be it is rambling, it is visually complete with plaintext image Difference, and therefrom cannot get any information image-related with plaintext.

Three, attacked by noise

The color Lena image that size is 512 × 512 is used as plaintext image with colour Baboon image by this experiment, is such as schemed (a), (b) are shown in 5, and the colored Lake image that size is 512 × 512 is as color host image, in Fig. 5 shown in (c), phase In the visual security ciphertext image such as Fig. 5 answered shown in (e), other parameters are constant.Visual security ciphertext image is distinguished in this experiment The salt-pepper noise (SPN) for adding varying strength, is then gone to decrypt noise-containing image using correct key, be solved accordingly Close image.

As shown in Figure 8.In fig. 8, (a) is the visual security image for not adding noise；(b)~(d) adds intensity respectively For the visual security image after 0.00003%, 0.00005%, 0.00007% salt-pepper noise (SPN)；(e) when being noiseless Decrypted image；(f)~(h) is after adding intensity as 0.00003%, 0.00005%, 0.00007% salt-pepper noise (SPN) Under decrypted image.

As can be seen from Figure 8: (1): after the salt-pepper noise of the ciphertext image addition varying strength of visual security, decryption figure Picture and decrypted image when not adding noise visually see difference very little；(2): with the increase of addition noise intensity, accordingly The visual quality of decrypted image also decline, but still can therefrom find out the information of plaintext image.

Four, shearing attack

The ciphertext image of visual security may partially or completely be lost in transmission storage, duplication are pasted.Therefore, have Necessity carries out shearing attack analysis to the compression of images Encryption Algorithm of proposition.In following emulation, this experiment is to (e) in Fig. 5 Shown in the visual security ciphertext image ability being sheared to examine the anti-shearing attack of the algorithm, other parameters are constant, survey Test result is as shown in Figure 9 and Figure 10.

In Fig. 9: (a) being visual security ciphertext；It (b) is decrypted image 1；It (c) is decrypted image 2.

In Figure 10: (a) being visual security ciphertext；It (b) is decrypted image 1；It (c) is decrypted image 2.

It can be seen that after ciphertext image is reduced a part from Fig. 9 and Figure 10, the image of decryption restoration still retains The important information that plaintext image is included, this illustrates that this algorithm has certain anti-shearing attacking ability.

Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although Present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: it still may be used To modify the technical solutions described in the foregoing embodiments or equivalent replacement of some of the technical features； And these are modified or replaceed, technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution spirit and Range.

Claims

1. the two width color image compression encryption methods based on two dimensional compaction perception characterized by comprising

Step 1: extracting the first width colour plaintext image P respectively₁With the second width colour plaintext image P₂Color three-component, be denoted as First group of color three-component and second group of color three-component, and using two-dimensional discrete wavelet conversion respectively to two groups of color three-components Carry out rarefaction；Wherein, P₁And P₂Size be M × N and M=N；

Step 2: calculating separately to obtain according to two groups of color three-components of two width colour plaintext images colored in plain text with two width The relevant first group information parameter of image and the second group information parameter；

Step 3: generating calculation matrix using LSS chaos system and Kronecker product, and the calculation matrix is optimized；

Step 4: two dimensional compaction perception is carried out to two groups of color three-components after rarefaction respectively using the calculation matrix after optimization, Obtain corresponding first group of compressed sensing measurement sequence and second group of compressed sensing measurement sequence；

Step 5: shuffle operation being carried out to two groups of compressed sensing measurement sequences respectively using sextuple hyperchaotic system, is obtained corresponding First group of Scrambling Matrix and second group of Scrambling Matrix, and the element value by element value in two groups of Scrambling Matrixes less than preset threshold TS It is changed to 0, obtains revised two groups of Scrambling Matrixes；

Step 6: revised two groups of Scrambling Matrixes being quantified respectively, element value is transformed between 0 to 255, are obtained pair The first group of quantization matrix and second group of quantization matrix answered；

Step 7: the color three-component of the color host image F ' after extracting image normalization is denoted as third group color three-component； Wherein, the size of color host image F ' is M × N；

Step 8: generating 256 cryptographic Hash using the SHA256 function of the first width colour plaintext image, and by 256 Hash It is worth every 8 one group and is converted into 32 decimal number k₁, k₂..., k₃₂, 32 decimal numbers are denoted as cryptographic Hash K；

Step 9: by the first group information parameter relevant to two width colour plaintext images and the second group information parameter according to setting Parameter Switch rule is respectively converted into first group of integer set and second group of integer set, wherein first group of integer set It include L integer with second group of integer set；

Step 10: (32+2L) a pixel value before one of component in third group color three-component is replaced with into the Kazakhstan Uncommon value K and 2L integer, obtain the first ciphertext matrix containing key parameter information；

Step 11: other two component in third group color three-component is subjected to DWT decomposition, and the matrix that will be obtained after decomposition Piecemeal is carried out, corresponding first group of carrier matrix and second group of carrier matrix are obtained；

Step 12: first group of quantization matrix and second group of quantization matrix being respectively embedded in two groups of carrier matrixs, obtained The second ciphertext matrix and third ciphertext matrix containing ciphertext image information；

Step 13: first ciphertext matrix, the second ciphertext matrix and third ciphertext matrix being combined, obtained and two width coloured silk The corresponding visual security image of color plaintext image.

2. the method according to claim 1, wherein further include: generate sextuple hyperchaotic system and LSS chaos system The initial value of system；Wherein:

Step A2: according to the intermediate parameters H₁~H₆Wherein three initial values of sextuple hyperchaotic system are calculated according to formula (10) x₀、y₀And z₀:

Wherein, mod (a, b) indicates a to the modulo operation of b, and abs (x) indicates to seek the absolute value of x, and floor (x) indicates to calculate not Maximum integer greater than x；

Step A3: according to the first group information parameter and the intermediate parameters H₁~H₆It is calculated among first group according to formula (11) Key r₁、g₁And b₁:

Wherein, Q_R、Q_GAnd Q_BFor the first group information parameter；

Step A4: according to the intermediate key r₁、g₁And b₁And intermediate parameters H₄~H₆Sextuple hyperchaos are calculated according to formula (12) The other three initial value u of system₀、v₀And w₀:

The initial value for generating LSS chaos system specifically includes:

Step B1: according to the second group information parameter, the intermediate parameters H₁~H₆With preset-key parameter t₁~t₅According to formula (13) second group of intermediate key r ' is calculated₁、g′₁With b '₁:

Wherein, Q '_R、Q′_GWith Q '_BFor the second group information parameter；

Step B2: according to second group of intermediate key r '₁、g′₁With b '₁The initial value of LSS chaos system is calculated according to formula (14) a₀With parameter r ':

Wherein, (t₆,t₇) ∈ (0,1) be preset-key parameter.

3. method according to claim 1 or 2, which is characterized in that the step 2 specifically:

Wherein, a (i, j) indicates (i, j) a element in corresponding with information parameter Q to be calculated color component, i.e., when a (i, J) color component R is respectively indicated₁、G₁、B₁、R₂、G₂、B₂In (i, j) a element when, Q respectively indicates corresponding information parameter Q_R、Q_G、Q_B、Q′_R、Q′_G、Q′_B；Wherein, R₁、G₁And B₁For first group of color three-component, R₂、G₂And B₂Divide for second group of color three Amount；Q_R、Q_GAnd Q_BFor the first group information parameter, Q '_R、Q′_GWith Q '_BFor the second group information parameter.

4. according to the method described in claim 2, it is characterized in that, utilizing in LSS chaos system and Crow in the step 3 Gram product generate calculation matrix specifically include:

Step 3.1: by the initial value a₀Iteration m in LSS chaos system is brought into parameter r '₁+ Md times, cast out preceding m₁A value, obtains The sequence A for being 1 × Md to length；Wherein, d is default sampling interval, m₁≥500；

A′_i=1-2A_i, i=1,2 ..., Md (15)

A″_k=A '_1+kdK is sampling sequence number, k >=0 (16)

Step 3.4: the chaos sequence S that 4 length are 4 is randomly selected from sequence A "₁、S₂、S₃And S₄, and by chaos sequence S₁~ S₄It is converted into the low-dimensional matrix K that 4 sizes are 2 × 2₁、K₂、K₃And K₄；

Step 3.5: using Kronecker product to low-dimensional matrix K₁And K₂、K₁And K₃、K₂And K₃And K₁And K₄Log is extended respectively₂ ^M Higher dimensional matrix K that is secondary, will obtaining₁′、K₂′、K₃' and K₄' it is used as calculation matrix.

5. according to the method described in claim 4, it is characterized in that, being optimized to the calculation matrix in the step 3 It specifically includes:

Step 3.6: by calculation matrix K₁' it is converted into calculation matrix φ, the φ ∈ R that size is P × Q^P×Q；Wherein, P=CR × M, Q =N, CR are default compression ratio；

Step 3.8: all 1's matrix J that construction size is M × N；Calculate the mean value var of diagonal matrix Σ diagonal entry₁, and count Σ₁In >=var₁Singular value number f；The preceding f of order matrix J is arranged multiplied by weighting coefficient t, obtains optimization matrix J₁, wherein t > 1； Matrix J will be optimized₁With calculation matrix φ dot product, calculation matrix φ ' after obtaining the first suboptimization；

Step 3.9: singular value decomposition, i.e. φ '=U are carried out again to calculation matrix φ '₁Σ₂ V₁ ^T；Wherein,Table Show the diagonal matrix of calculation matrix φ ', Σ₃=diag (δ '₁,δ′₂..., δ '_r),δ′₁~δ '_rIndicate the surprise of calculation matrix φ ' Different value, r are singular value number；

Step 3.10: enabling δ '₁=δ '₂=...=δ '_r=1, obtain new diagonal matrix Σ '₂, and according to diagonal matrix Σ '₂It is raw At the calculation matrix φ after final optimization pass₁=U₁Σ′₂V₁；

Step 3.11: referring to step 3.6 to step 3.10, to calculation matrix K₂′、K₃' and K₄' optimize, after being optimized Calculation matrix φ₂、φ₃And φ₄。

6. according to the method described in claim 2, it is characterized in that, the step 5 specifically includes:

Step 5.1: by the initial value x₀、y₀、z₀、u₀、v₀And w₀Bring iteration m in sextuple hyperchaotic system into₀+M₁N₁It is secondary, cast out Preceding m₀A value, obtaining 6 sizes is 1 × M₁N₁Sequence X, Y, Z, U, V and W；Wherein, m₀>=500, M₁=M × CR, N₁=N × CR, CR are default compression ratio；

Step 5.3: six chaos sequences X, Y, Z, U, V and W being modified according to formula (19), obtain revised new sequence XX, YY:

Wherein, XX_(i)、YY_(i)Indicate i-th of element in new chaos sequence XX and YY, X_i、Y_i、Z_i、U_i、V_iAnd W_iIt respectively indicates I-th of element of sequence X, Y, Z, U, V and W, i ∈ [1, M₁N₁]；

Step 5.4: extracting first element value X of new sequence X X, YY respectively_ind1、Y_ind1, and be modified according to formula (20), it obtains To revised sequence X '_ind1With Y '_ind1:

Step 5.5: if X '_ind1=1, index of reference vector D₁、D₂And D₃Respectively to sequence R₃、G₃And B₃Rearrangement；If X′_ind1=2, index of reference vector D₁、D₂And D₃Respectively to sequence G₃、B₃And R₃Rearrangement；If X '_ind1=3, index of reference vector D₁、D₂And D₃To sequence B₃、R₃And G₃Rearrangement；Wherein, R₃、G₃And B₃Sequence is measured for first group of compressed sensing；

Step 5.6: if Y '_ind1=1, index of reference vector D₄、D₅And D₆To sequence R '₃、G′₃With B '₃Rearrangement；If Y '_ind1 =2, index of reference vector D₄、D₅And D₆To sequence G '₃、B′₃With R '₃Rearrangement；If Y '_ind1=3, index of reference vector D₄、D₅With D₆To sequence B '₃、R′₃With G '₃Rearrangement；Wherein, R '₃、G′₃With B '₃Sequence is measured for second group of compressed sensing；

Step 5.7: first group of compressed sensing after scramble is measured into sequence R₃、G₃And B₃And second group of compressed sensing measures sequence Arrange R '₃、G′₃With B '₃Being separately converted to size is M₁×N₁Matrix R₄、G₄And B₄And R '₄、G′₄With B '₄, wherein R₄、G₄ And B₄For first group of Scrambling Matrix, R '₄、G′₄With B '₄For second group of Scrambling Matrix.

7. according to the method described in claim 6, it is characterized in that, the step 6 is specially to be quantified according to formula (21):

Wherein, min and max is the minimum value and maximum value of Scrambling Matrix P respectively, and floor (x) indicates to calculate no more than x most Big integer；P_(i)With P '_(i)It is i-th of element of Scrambling Matrix P and quantization matrix P ' respectively, that is, works as P_(i)Respectively indicate scramble square Battle array R₄、G₄、B₄、R′₄、G′₄、B′₄In i-th of element when, P '_(i)It is quantization matrix R respectively₅、G₅、B₅、R′₅、G′₅、B′₅I-th A element；1≤i≤M₁N₁；Wherein, R₅、G₅And B₅For first group of quantization matrix；R′₅、G′₅With B '₅For second group of quantization matrix.

8. the method according to claim 1, wherein the step 11 specifically:

Step 11.1: DWT decomposition being carried out to other two component in third group color three-component, respectively obtaining size is (M × N)/4 Matrix C_A、C_H、C_V、C_DWith C '_A、C′_H、C′_V、C′_D；Wherein, other two component in third group color three-component Size is M × N；

Step 11.3: by Matrix C_HFour pieces are divided into, 4 matrix L L that size is (M × N)/16 are obtained₁、LH₁、HL₁And HH₁；

Step 11.4: referring to step 11.3, by Matrix C_V、C′_H、C′_VIt is divided into four pieces respectively, obtaining size is (M × N)/16 Other 12 matrix L L₂~LL₄、LH₂~LH₄、HL₂~HL₄And HH₂~HH₄；Wherein, LL₁~LL₂、LH₁~LH₂、HL₁~HL₂、 HH₁~HH₂For first group of carrier matrix；LL₃~LL₄、LH₃~LH₄、HL₃~HL₄And HH₃~HH₄For second group of carrier matrix.

9. according to the method described in claim 8, it is characterized in that, the step 12 specifically:

Step 12.1: two groups of quantization matrixes being converted according to formula (22) and formula (23) respectively, every group of quantization matrix obtains To transformed two matrixes d and d':

D (i, j)=Q (i, j) mod10 (22)

D ' (i, j)=floor (Q (i, j)/10) (23)

Wherein, Q (i, j) and d (i, j), d ' (i, j) are matrix d, the d ' obtained after quantization matrix Q and Q to be transformed are converted respectively In (i, j) a element, i.e., when Q (i, j) is quantization matrix R respectively₅、G₅、B₅、R′₅、G′₅、B′₅In (i, j) a element When, d (i, j) is corresponding d respectively₁、d₂、d₃、d₄、d₅、d₆In (i, j) a element, d ' (i, j) is corresponding d respectively₁₁、 d₂₂、d₃₃、d₄₄、d₅₅、d₆₆In (i, j) a element；1≤i≤M₁, 1≤j≤N₁；

Wherein, C (i, j) is (i, j) a element in Matrix C to be calculated, and I is the pixel average of Matrix C to be calculated, that is, is worked as C (i, j) is C respectively_AWith C '_AIn (i, j) a element when, I is C respectively_AWith C '_APixel average I₁And I₂；

Step 12.3: extracting Matrix C_AWith C '_AFirst element value E₁And E₂；

Step 12.4: being respectively compared I₁And E₁Size and I₂And E₂Size, and according to comparison result by d₁~d₆And d₁₁~ d₆₆It is embedded into first group of carrier matrix and second group of carrier matrix, first group of carrier matrix LL ' after being embedded in₁~ LL′₂、LH′₁~LH '₂、HL′₁~HL '₂、HH′₁~HH '₂, and second group of carrier matrix LL ' after insertion₃~LL '₄、LH′₃ ~LH '₄、HL′₃~HL '₄、HH′₃~HH '₄；

Step 12.5: the matrix L L ' that will be obtained after insertion₁、LH′₁、HL′₁、HH′₁It is reconfigured, obtains new Matrix C_H1；

Step 12.6: referring to step 12.5, the matrix L L ' that will be obtained after insertion₂~LL '₄、LH′₂~LH '₄、HL′₂~HL '₄With HH′₂~HH '₄It is reconfigured, obtains new Matrix C_V1、C′_H1With C '_V1；

Step 12.7: by Matrix C_A、C_H1、C_V1、C_DWith C '_A、C′_H1、C′_V1、C′_DThe second ciphertext is obtained using inverse discrete wavelet transform Matrix and third ciphertext matrix.

10. according to the method described in claim 9, it is characterized in that, the step 12.4 specifically:

If E₁≥I₁And E₂≥I₂, d₁It is embedded into LL₁In, d₁₁It is embedded into LH₁In, d₂It is embedded into HL₁In, d₂₂It is embedded into HH₁In, d₃It is embedded into LL₂In, d₃₃It is embedded into LH₂In, d₄It is embedded into LL₃In, d₄₄It is embedded into LH₃In, d₅Insertion To HL₃In, d₅₅It is embedded into HH₃In, d₆It is embedded into HL₄In, d₆₆It is embedded into HH₄In；

If E₁≥I₁And E₂<I₂, d₁It is embedded into LL₂In, d₁₁It is embedded into LH₂In, d₂It is embedded into LL₃In, d₂₂It is embedded into LH₃In, d₃It is embedded into HL₃In, d₃₃It is embedded into HH₃In, d₄It is embedded into HL₄In, d₄₄It is embedded into HH₄In, d₅Insertion To LL₁In, d₅₅It is embedded into LH₁In, d₆It is embedded into HL₁In, d₆₆It is embedded into HH₁In；

If E₁<I₁And E₂≥I₂, d₁It is embedded into LL₃In, d₁₁It is embedded into LH₃In, d₂It is embedded into HL₃In, d₂₂It is embedded into HH₃In, d₃It is embedded into HL₄In, d₃₃It is embedded into HH₄In, d₄It is embedded into LL₁In, d₄₄It is embedded into LH₁In, d₅Insertion To HL₁In, d₅₅It is embedded into HH₁In, d₆It is embedded into LL₂In, d₆₆It is embedded into LH₂In；

If E₁<I₁And E₂<I₂, d₁It is embedded into HL₄In, d₁₁It is embedded into HH₄In, d₂It is embedded into LL₁In, d₂₂It is embedded into LH₁ In, d₃It is embedded into HL₁In, d₃₃It is embedded into HH₁In, d₄It is embedded into LL₂In, d₄₄It is embedded into LH₂In, d₅It is embedded into LL₃In, d₅₅It is embedded into LH₃In, d₆It is embedded into HL₃In, d₆₆It is embedded into HH₃In.