Image encryption method based on inner and outer channel permutation and Fibonacci variable matrix

Abstract

The application discloses an image encryption method based on inner and outer channel permutation and Fibonacci variable matrix, adopts a two-dimensional cosine polynomial hyperchaotic mapping with a wider chaotic parameter range and higher randomness, and compared with an existing two-dimensional chaotic system, the Lyapunov exponent is significantly improved, and the problems of limited chaotic performance and insufficient randomness can be effectively solved. Through the combination of three-channel grouping and inner and outer scrambling algorithms, the correlation between pixels is effectively destroyed, and the anti-statistical analysis ability of the image is significantly improved. Meanwhile, based on the chaotic variable matrix and the pseudo-random translation vector design of the Fibonacci sequence, the diffusion process exhibits a strong avalanche effect, so that the image shows stronger resistance to difference attack and selected plaintext attack. The low-complexity inner and outer channel permutation algorithm and the two-dimensional matrix diffusion method can reduce the calculation overhead, and compared with a traditional algorithm, the execution efficiency is significantly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image encryption, and in particular to an image encryption method based on internal and external channel permutation and Fibonacci variable matrix. Background Art

[0002] With the rapid development of medical cloud services and remote consultation technologies, unauthorized access to patient information poses a serious threat to information security. In medical scenarios, preventing the leakage or tampering of patient personal information has become a critical issue that needs to be addressed. Furthermore, to meet the low-latency requirements of information transmission, the development of fast encryption algorithms is essential. Image encryption, as the most direct and secure method for protecting image information, has gradually attracted widespread attention due to its ability to effectively prevent information leakage. However, due to the large amount of data and high correlation between pixels in images, traditional encryption methods (such as DES and AES) face significant challenges in image encryption. Chaotic image encryption systems, due to their high sensitivity to initial conditions, long periodicity, and unpredictable iteration trajectories, have become a more suitable solution for image encryption.

[0003] Existing chaotic systems are primarily categorized as one-dimensional and higher-dimensional. One-dimensional chaotic systems have a simple structure, but their chaotic trajectories are vulnerable to attack, resulting in lower security. In contrast, higher-dimensional chaotic systems, with their more complex structures and enhanced chaotic properties, are a hot research area. However, increasing the dimensionality of a chaotic system significantly increases computational complexity, leading to the widespread adoption of two-dimensional chaotic systems as a trade-off. Two-dimensional chaotic systems offer a certain balance between computational efficiency and chaotic properties, making them an important area of ​​research in image encryption.

[0004] Furthermore, Fridrich's permutation-diffusion image encryption method based on two-dimensional chaotic mapping is considered a typical architecture and has been widely used in image encryption. In recent years, chaos-based image encryption methods have further integrated techniques such as DNA coding, neural networks, and group permutation. Group permutation schemes, among others, are particularly well-suited for fast encryption due to their ability to quickly perform scrambling and effectively destroy inter-pixel correlations.

[0005] However, the above-mentioned prior art all has the following shortcomings:

[0006] 1. Existing two-dimensional chaotic systems (such as 2D-SPHM, 2D-SSCDB, and Cross-2DHM) have a narrow chaotic parameter range and insufficient randomness, making it difficult to effectively destroy inter-pixel correlations. These systems are particularly difficult to meet in high-security and high-efficiency application scenarios.

[0007] 2. Regarding channel group scrambling, existing methods (such as the three-channel scrambling method based on the Arnold map) are limited by the periodicity of chaotic attractors, which affects their security. Furthermore, while the Fisher-Yates scrambling algorithm can achieve effective intra-block scrambling, its high time complexity makes it unsuitable for fast encryption applications.

[0008] 3. Although existing diffusion algorithms based on linear matrix transformation (such as the Hill cipher) are widely used due to their simplicity, speed, and effectiveness, the key matrices of their original and improved versions must be reversible in ring Z / 256Z, otherwise decryption will fail, increasing the complexity and risk of their applications.

[0009] 4. As most images transmitted over the Internet are color images, encryption of the RGB channels has become a research focus. However, existing methods struggle to achieve both efficiency and security in terms of three-channel grouping and scrambling.

[0010] 5. Many existing algorithms fail to meet ideal execution times. For example, some encryption schemes based on DNA coding or complex chaotic systems have significant shortcomings in image processing speed, making it difficult to meet the requirements of real-time encryption. Summary of the Invention

[0011] The purpose of the present invention is to overcome the deficiencies of the prior art and provide an image encryption method based on inner and outer channel permutation and Fibonacci matrix transformation.

[0012] To achieve the above objectives, the technical solutions provided by the present invention are:

[0013] An image encryption method based on internal and external channel permutation and Fibonacci matrix transformation, comprising:

[0014] Input the plaintext image P, and perform scrambling processing on the plaintext image P based on the two-dimensional cosine polynomial hyperchaotic mapping system and the internal and external channel permutation algorithm to obtain the scrambled image P′;

[0015] The model of the two-dimensional cosine polynomial hyperchaotic mapping system is as follows:

[0016]

[0017] y v+1 =b·cos(ay v -bx v +x v )

[0018] Among them, x v and y vis the input variable of the chaotic mapping; v is the number of iterations of the two-dimensional chaotic system; a and b are the control parameters of the mapping, which are used to adjust the chaotic behavior of the system;

[0019] The scrambled image P′ is diffused by the modulo-π Fibonacci matrix diffusion algorithm, and finally the encrypted image is obtained.

[0020] Furthermore, the plaintext image P is scrambled based on the internal and external channel replacement algorithm, including:

[0021] A1. Use a two-dimensional cosine polynomial hyperchaotic mapping system to generate three chaotic matrices C, Q, and S. These chaotic matrices are adjusted to the same size as the input plaintext image P. The elements of the chaotic matrix C range from [1, 6], and the elements of the chaotic matrix Q range from [1, M]. M is the length of the plaintext image P.

[0022] A2. For each pixel position R(i,j)G(i,j)B(i,j), use the value of the chaotic matrix C(i,j) as the index and select the corresponding three-channel arrangement order from the preset channel sequence list selectedlist;

[0023] Selected_Channels(i,j)=selected_list[C(i,j)]

[0024] A3. Combine Selected_Channels(i,j) and the chaotic matrix Q to obtain the channel image grouping matrix G;

[0025] A4. Group the image grouping matrix G by rows, and in combination with the chaotic matrix S, scramble each group using the internal and external channel scrambling algorithm to achieve element scrambling exchange;

[0026] A5. Repeat step A4 until all elements in each group are repositioned, and finally obtain the scrambled image P′.

[0027] Furthermore, combined with Selected_Channels(i,j), the channel image grouping matrix G is obtained, including:

[0028] The value of each pixel in the image grouping matrix G is determined by a triplet, in which:

[0029] The first component represents the row index of the retrieved value from the original image, which is equal to the value of the chaotic matrix Q at the pixel position R(i,j)G(i,j)B(i,j);

[0030] The second component represents the column index, which is equal to the column index of the element position R(i,j)G(i,j)B(i,j);

[0031] The third component represents the channel index, which is equal to the sequence value of the current channel Selected_Channels(i,j);

[0032] Based on the triplet, the value of the corresponding position is extracted from the plaintext image P and filled into the corresponding position of the image grouping matrix G, thereby obtaining the channel image grouping matrix G.

[0033] Furthermore, step A4 includes:

[0034] The image grouping matrix G is grouped by rows, and each group is scrambled by the internal and external channel scrambling algorithm:

[0035] One row is a group, and the kth row is grouped into G k , whose size is 1×N, where N is the width of the plaintext image P, namely:

[0036] G k ={g k0 , g k1 ,...,g kN-1}.

[0037] Among them, g kN-1 Indicates group G k The kN-1th element in ;

[0038] Group the chaotic matrix S by rows, and the corresponding chaotic sequence is:

[0039] S k ={s k0 , s k1 ,...,s kN-1}

[0040] Among them, s kN-1 represents the kN-1th element in the kth row of the chaotic matrix S;

[0041] The traversal pointer starts from the starting position of the subgroup and moves in sequence after each replacement iteration; the chaotic sequence value of the current round and the position of the traversal pointer determine the replacement pointer index value, starting from the position of the traversal pointer and continuing to the position pointed by the chaotic sequence value;

[0042] Scrambling index S shuffle (n) is defined as:

[0043]

[0044] Among them, n is the position of the current traversal pointer, s kn is the chaotic sequence S k The corresponding element in; This formula determines the new index position of the replacement pointer according to the chaotic sequence;

[0045] The scrambling process is as follows:

[0046]

[0047] in, Indicates element exchange, when the traversal pointer traverses the group G k After all elements of , the scrambling of the group is completed.

[0048] Furthermore, the scrambled image P′ is diffused by using the modulo-π Fibonacci matrix diffusion algorithm, including:

[0049] The two-dimensional cosine polynomial hyperchaotic mapping system is used to perform M×N iterations, where M is the length of the plaintext image P and N is the width of the plaintext image P, to generate three chaotic sequences K1, K2 and K3.

[0050] Generate a Fibonacci sequence and apply the modular π operation to obtain a Fibonacci sequence variant with chaotic effect, which is used to generate the key matrix parameter f 11 ,f 12 ,f 21 ,f 22 , and the key matrix parameter f 11 Reversible in ring Z or 256Z;

[0051] Read the scrambled image P′ and obtain its size M×N;

[0052] Convert the scrambled image P′ into vector form;

[0053] The first pixel P′(1) is encrypted and its adjacent pixels P′(2) are scrambled as follows:

[0054]

[0055] B(2)=(Z(1)+P′(2))mod256

[0056] Where A(1) is the first encrypted pixel, P′(1) is the first pixel value of the scrambled image, B(2) is the second scrambled pixel, and Z(1) is the temporary value used to randomize adjacent pixels.

[0057] The remaining pixels are encrypted as follows:

[0058]

[0059] B(m)=(Z(m-1)+P′m))mod256

[0060] Where m represents the index of the current pixel;

[0061] Extend the encryption to the last pixel, thus completing the encryption process of the entire image;

[0062] Finally, the vector A is converted back into a two-dimensional matrix form to obtain the encrypted image.

[0063] Compared with the existing technology, the principles and advantages of this technical solution are as follows:

[0064] 1. A two-dimensional cosine polynomial hyperchaotic mapping (2D-CPHM) with a wider chaotic parameter range and higher randomness is adopted. Compared with existing two-dimensional chaotic systems (such as 2D-SPHM, 2D-SSCDB, and Cross-2DHM), its Lyapunov exponents (LE1=222.1670, LE2=256.6653) are significantly improved, which can effectively solve the problems of limited chaotic performance and insufficient randomness.

[0065] 2. By combining three-channel grouping with internal and external scrambling algorithms, the correlation between pixels is effectively destroyed, significantly improving the image's resistance to statistical analysis. Furthermore, based on the chaotic variable matrix and pseudo-random translation vector design of the Fibonacci sequence, the diffusion process exhibits a powerful avalanche effect, making the image more resistant to differential attacks and chosen-plaintext attacks.

[0066] 3. The use of a low-complexity inner and outer channel permutation algorithm and a two-dimensional matrix diffusion method reduces computational overhead and significantly improves execution efficiency compared to traditional algorithms. For example, in a 512×512 image test, the algorithm encryption time was only 0.32 seconds, significantly outperforming existing algorithms and meeting the requirements of low-latency applications.

[0067] 4. It supports multi-channel encryption of color images and is applicable to image data of any shape and size. It has good versatility and adaptability and is particularly suitable for scenarios such as medical cloud services and remote consultations that have high requirements for information security and encryption speed.

[0068] 5. Using multiple chaotic initial parameters and Fibonacci sequence initial values ​​to construct the key, the key space reaches 2^520, making it more resistant to brute force attacks than existing algorithms. Furthermore, the system's high sensitivity to the key ensures that even minor changes in the key will result in incorrect decryption, further enhancing security. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the services required for use in the embodiments or the prior art descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0070] Figure 1 This is a principle flow chart of an image encryption method based on internal and external channel permutation and Fibonacci matrix transformation according to the present invention;

[0071] Figure 2 Bifurcation diagrams of the two-dimensional cosine polynomial hyperchaotic mapping (2D-CPHM) used in the present invention under different parameters ((a1) and (b1) are two-dimensional diagrams (b=25); (d1) and (e1) are two-dimensional diagrams (a=25); (c1) and (f1) are three-dimensional diagrams);

[0072] Figure 3 The two-dimensional and three-dimensional phase space trajectory diagrams of the two-dimensional cosine polynomial hyperchaotic mapping (2D-CPHM) ((a2) two-dimensional phase space trajectory with a=25, b=25; (b2) three-dimensional phase space trajectory with a=25; (c2) three-dimensional phase space trajectory with b=25);

[0073] Figure 4 Schematic diagram of the three-dimensional Lyapunov exponent of different two-dimensional chaotic maps;

[0074] Figure 5 Schematic diagram of the two-dimensional Lyapunov exponent of different two-dimensional chaotic maps;

[0075] Figure 6 It is the original image (plaintext image) used in the encrypted image test;

[0076] Figure 7 This is the encrypted image in the encrypted image test;

[0077] Figure 8 A schematic diagram that illustrates the sensitivity of the key. DETAILED DESCRIPTION

[0078] The present invention will be further described below in conjunction with specific embodiments:

[0079] like Figure 1 As shown, the image encryption method based on inner and outer channel permutation and Fibonacci matrix transformation described in this embodiment includes the following steps:

[0080] S1. Input a plaintext image P and perform scrambling on the plaintext image P based on a two-dimensional cosine polynomial hyperchaotic mapping system and an internal and external channel permutation algorithm to obtain a scrambled image P'.

[0081] The model of the two-dimensional cosine polynomial hyperchaotic mapping system is as follows:

[0082]

[0083] y v+1 =b·cos(ayv -bx v +x v )

[0084] Among them, x v and y v is the input variable of the chaotic mapping; v is the number of iterations of the two-dimensional chaotic system; a and b are the control parameters of the mapping, which are used to adjust the chaotic behavior of the system;

[0085] In this step, the plaintext image P is scrambled based on the internal and external channel replacement algorithm, including:

[0086] A1. Use a two-dimensional cosine polynomial hyperchaotic mapping system to generate three chaotic matrices C, Q, and S. These chaotic matrices are adjusted to the same size as the input plaintext image P. The elements of the chaotic matrix C range from [1, 6], and the elements of the chaotic matrix Q range from [1, M]. M is the length of the plaintext image P.

[0087] A2. For each pixel position R(i,j)G(i,j)B(i,j), use the value of the chaotic matrix C(i,j) as the index and select the corresponding three-channel arrangement order from the preset channel sequence list selectedlist;

[0088] Selected_Channels(i,j)=selected_list[C(i,j)]

[0089] Specifically, selectedlist is a preset channel sequence list generated by enumerating 6 permutations and combinations of RGB channels, which is used for indexing the chaotic matrix C;

[0090] A3. Combine Selected_Channels(i,j) and the chaotic matrix Q to obtain the channel image grouping matrix G;

[0091] This sub-step specifically includes:

[0092] The value of each pixel in the image grouping matrix G is determined by a triplet, in which:

[0093] The first component represents the row index of the retrieved value from the original image, which is equal to the value of the chaotic matrix Q at the pixel position R(i,j)G(i,j)B(i,j);

[0094] The second component represents the column index, which is equal to the column index of the element position R(i,j)G(i,j)B(i,j);

[0095] The third component represents the channel index, which is equal to the sequence value of the current channel Selected_Channels(i,j);

[0096] Based on the triplet, the value of the corresponding position is extracted from the plaintext image P and filled into the corresponding position of the image grouping matrix G, thereby obtaining the channel image grouping matrix G.

[0097] A4. Group the image grouping matrix G by rows, and in combination with the chaotic matrix S, scramble each group using the internal and external channel scrambling algorithm to achieve element scrambling exchange;

[0098] This sub-step specifically includes:

[0099] The image grouping matrix G is grouped by rows, and each group is scrambled by the internal and external channel scrambling algorithm:

[0100] One row is a group, and the kth row is grouped into G k , whose size is 1×N, where N is the width of the plaintext image P, namely:

[0101] G k ={g k0 , g k1 ,...,g kN-1}

[0102] Among them, g kN-1 Indicates group G k The kN-1th element in ;

[0103] Group the chaotic matrix S by rows, and the corresponding chaotic sequence is:

[0104] S k ={s k0 , s k1 ,...,s kN-1}

[0105] Among them, s kN-1 represents the kN-1th element in the kth row of the chaotic matrix S;

[0106] The traversal pointer starts from the starting position of the subgroup and moves in sequence after each replacement iteration; the chaotic sequence value of the current round and the position of the traversal pointer determine the replacement pointer index value, starting from the position of the traversal pointer and continuing to the position pointed by the chaotic sequence value;

[0107] Scrambling index S shuffle (n) is defined as:

[0108]

[0109] Among them, n is the position of the current traversal pointer, s kn is the chaotic sequence S k The corresponding element in; This formula determines the new index position of the replacement pointer according to the chaotic sequence;

[0110] The scrambling process is as follows:

[0111]

[0112] in, Indicates element exchange, when the traversal pointer traverses the group G k After all elements of , the scrambling of the group is completed.

[0113] A5. Repeat step A4 until all elements in each group are repositioned, and finally obtain the scrambled image P'.

[0114] S2. Diffusion processing is performed on the scrambled image P′ through the modulo-π Fibonacci matrix diffusion algorithm to obtain an encrypted image.

[0115] This step specifically includes:

[0116] The two-dimensional cosine polynomial hyperchaotic mapping system is used to perform M×N iterations, where M is the length of the plaintext image P and N is the width of the plaintext image P, to generate three chaotic sequences K1, K2 and K3.

[0117] Generate a Fibonacci sequence and apply the modular π operation to obtain a Fibonacci sequence variant with chaotic effect, which is used to generate the key matrix parameter f 11 ,f 12 ,f 21 ,f 22 , and the key matrix parameter f 11 Reversible in ring Z or 256Z;

[0118] Read the scrambled image P′ and obtain its size M×N;

[0119] Convert the scrambled image P′ into vector form;

[0120] The first pixel P'(1) is encrypted and its adjacent pixels P'(2) are scrambled as follows:

[0121]

[0122] B(2)=(Z(1)+P′(2))mod256

[0123] Where A(1) is the first encrypted pixel, P′(1) is the first pixel value of the scrambled image, B(2) is the second scrambled pixel, and Z(1) is the temporary value used to randomize adjacent pixels.

[0124] The remaining pixels are encrypted as follows:

[0125]

[0126] B(m)=(Z(m-1)+P′m))mod256

[0127] Where m represents the index of the current pixel;

[0128] Extend the encryption to the last pixel, thus completing the encryption process of the entire image;

[0129] Finally, the vector A is converted back into a two-dimensional matrix form to obtain the encrypted image.

[0130] In order to demonstrate the superiority of the method of the present invention, the following analytical experiments were performed:

[0131] 1. Analyze the two-dimensional cosine polynomial hyperchaotic mapping (2D-CPHM) used. Figure 2 The bifurcation diagrams of 2D-CPHM under different parameters are shown. The trajectories are evenly distributed and have no periodic behavior, indicating good ergodicity. Figure 3 The two-dimensional and three-dimensional phase space trajectories show that the trajectory distribution of 2D-CPHM is complex and uniform, further supporting its applicability as a chaotic encryption core.

[0132] 2. Perform Lyapunov index test:

[0133] Table 1 shows that the two Lyapunov exponents (LE1 and LE2) of the 2D-CPHM system are both positive, 222.1670 and 256.6653 respectively, indicating that the system has significant hyperchaotic characteristics.

[0134]

[0135] Table 1

[0136] Figure 4 The three-dimensional Lyapunov exponent distribution of 2D-CPHM is demonstrated, showing that it maintains a high level of chaotic properties over a wide range of parameters. Figure 5 It is further shown that the two-dimensional distribution of the Lyapunov exponent as a function of the control parameters proves that the chaos intensity and range of 2D-CPHM are superior to those of traditional chaotic systems.

[0137] 3. Encrypted image test

[0138] like Figure 6 and Figure 7 As shown, a3, b3, and c3 are all original images (plaintext images), and d3, e3, and f3 are encrypted images. It can be seen that the method of the present invention has a good image encryption effect.

[0139] 4. Encryption Time Analysis

[0140]

[0141]

[0142] Table 2 Execution time analysis of different images

[0143] Table 2 shows the superiority of the image encryption algorithm of the present invention in encryption time, proving its high encryption efficiency.

[0144] 5. Key space analysis

[0145] For an image of size 512, assuming the calculation accuracy is 10^16, the key space is: 512×(10^16×10^16×10^16)^2≈2^520. Compared with the existing algorithm, the present invention is more resistant to brute force cracking.

[0146] 6. Key sensitivity analysis

[0147]

[0148] Among them, SNR is signal-to-noise ratio, MSE is mean square error, PSNR is peak signal-to-noise ratio, and SSIM is structural similarity;

[0149] project <![CDATA[x1]]> <![CDATA[y1]]> a b <![CDATA[N0]]> Change value <![CDATA[10 -15 ]]> <![CDATA[10 -15 ]]> 1 2 1

[0150] Table 3

[0151]

[0152]

[0153] Table 4. Evaluation of encrypted image differences due to encryption key changes

[0154] Evaluation metrics such as SNR, MSE, PSNR, and SSIM are used to assess the difference between encrypted or decrypted images. The more significant the difference between the two images, the higher the MSE and the lower the SNR, PSNR, and SSIM. If the encryption scheme is insensitive to key changes, then a small modification to the key will not change the structural differences between the two encrypted images. Figure 8 In this experiment, a4 is the original image, b4, c4, d4, and e4 are the images decrypted using slightly modified keys, and f4 is the image decrypted using the correct key. This experiment demonstrates the key sensitivity of the encryption scheme of the present invention, showing that even a single bit change in the key effectively prevents any retrieval of meaningful information.

[0155] The embodiments described above are only preferred embodiments of the present invention and are not intended to limit the scope of implementation of the present invention. Therefore, any changes made based on the shape and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. An image encryption method based on internal and external channel permutation and Fibonacci matrix transformation, characterized in that: include: Input the plaintext image P, and perform scrambling processing on the plaintext image P based on the two-dimensional cosine polynomial hyperchaotic mapping system and the internal and external channel replacement algorithm to obtain the scrambled image P ' ; The model of the two-dimensional cosine polynomial hyperchaotic mapping system is as follows: y v+1 =b·cos(a v -bx v +x v ) Among them, x v and y v is the input variable of the chaotic mapping; v is the number of iterations of the two-dimensional chaotic system; a and b are the control parameters of the mapping, which are used to adjust the chaotic behavior of the system; The scrambled image P is transformed by the Fibonacci matrix diffusion algorithm ' Perform diffusion processing and finally obtain the encrypted image; The plaintext image P is scrambled based on the internal and external channel replacement algorithm, including: A1. Use a two-dimensional cosine polynomial hyperchaotic mapping system to generate three chaotic matrices C, Q, and S. These chaotic matrices are adjusted to the same size as the input plaintext image P. The elements of the chaotic matrix C range from [1, 6], and the elements of the chaotic matrix Q range from [1, M]. M is the length of the plaintext image P. A2. For each pixel position R(i,j)G(i,j)B(i,j), use the value of the chaotic matrix C(i,j) as the index and select the corresponding three-channel arrangement order from the preset channel sequence list selected list; Selected_Channels(i,j)=selected_list[C(i,j)] A3. Combine Selected_Channels(i,j) and the chaotic matrix Q to obtain the channel image grouping matrix G; A4. Group the image grouping matrix G by rows, and in combination with the chaotic matrix S, scramble each group using the internal and external channel scrambling algorithm to achieve element scrambling exchange; A5. Repeat step A4 until all elements in each group are repositioned, and finally obtain the scrambled image P ' .

2. The image encryption method based on internal and external channel permutation and Fibonacci matrix transformation according to claim 1 is characterized in that: Combined with Selected_Channels(i,j) to obtain the channel image grouping matrix G, including: The value of each pixel in the image grouping matrix G is determined by a triplet, in which: The first component represents the row index of the retrieved value from the original image, which is equal to the value of the chaotic matrix Q at the pixel position R(i,j)G(i,j)B(i,j); The second component represents the column index, which is equal to the column index of the element position R(i,j)G(i,j)B(i,j); The third component represents the channel index, which is equal to the sequence value of the current channel Selected_Channels(i,j); Based on the triplet, the value of the corresponding position is extracted from the plaintext image P and filled into the corresponding position of the image grouping matrix G, thereby obtaining the channel image grouping matrix G.

3. The image encryption method based on internal and external channel permutation and Fibonacci matrix transformation according to claim 1 is characterized in that: Step A4 includes: The image grouping matrix G is grouped by rows, and each group is scrambled by the internal and external channel scrambling algorithm: One row is a group, and the kth row is grouped into G k , whose size is 1×N, where N is the width of the plaintext image P, namely: G k ={g k0 ,g k1 ,...,g kN-1 } Among them, g kN-1 Indicates group G k The kN-1th element in ; Group the chaotic matrix S by rows, and the corresponding chaotic sequence is: S k ={s k0 ,s k1 ,...,s kN-1 } Among them, s kN-1 represents the kN-1th element in the kth row of the chaotic matrix S; The traversal pointer starts from the starting position of the subgroup and moves in sequence after each replacement iteration; the chaotic sequence value of the current round and the position of the traversal pointer determine the replacement pointer index value, starting from the position of the traversal pointer and continuing to the position pointed by the chaotic sequence value; Scrambling index S shuffle (n) is defined as: Among them, n is the position of the current traversal pointer, s kn is the chaotic sequence S k The corresponding element in; This formula determines the new index position of the replacement pointer according to the chaotic sequence; The scrambling process is as follows: in, Indicates element exchange, when the traversal pointer traverses the group G k After all elements of , the scrambling of the group is completed.

4. The image encryption method based on internal and external channel permutation and Fibonacci matrix transformation according to claim 1 is characterized in that: The scrambled image P is transformed by the Fibonacci matrix diffusion algorithm ' Perform diffusion treatment, including: The two-dimensional cosine polynomial hyperchaotic mapping system is used to perform M×N iterations, where M is the length of the plaintext image P and N is the width of the plaintext image P, to generate three chaotic sequences K1, K2 and K3. Generate a Fibonacci sequence and apply the modular π operation to obtain a Fibonacci sequence variant with chaotic effect, which is used to generate the key matrix parameter f 11 ,f 12 ,f 21 ,f 22 , and the key matrix parameter f 11 Reversible in ring Z or 256Z; Read the scrambled image P ' , get its size M×N; The scrambled image P ' Convert to vector form; For the first pixel P ' (1) Encrypt and encode its adjacent pixels P in the following way ' (2) Perform scrambling: B(2)=(Z(1)+P′(2))mod256 Among them, A(1) is the first encrypted pixel, P ' (1) is the first pixel value of the scrambled image, B(2) is the second scrambled pixel, and Z(1) is the temporary value used to randomize adjacent pixels; The remaining pixels are encrypted as follows: B(m)=(Z(m-1)+P′m))mod256 Where m represents the index of the current pixel; Extend the encryption to the last pixel, thus completing the encryption process of the entire image; Finally, the vector A is converted back into a two-dimensional matrix form to obtain the encrypted image.

Images