Image semantic feature based adaptive encryption compression and power multiplexing transmission system

Abstract

The application discloses an image semantic feature-based adaptive encryption compression and power multiplexing transmission system and relates to the technical field of image processing. The original sending image is subjected to image saliency detection, and the original image is split into a plurality of image feature blocks of different levels according to saliency. In combination with a key, a chaotic sequence is generated through a chaotic system to perform adaptive encryption compression on the image feature blocks of different importance levels. Then, the data information after the encryption compression is subjected to power domain multiplexing modulation, wherein the data information of different importance levels is allocated with different powers to guarantee the transmission performance of the key data information. After transmission through an optical fiber channel, the receiving end completes image feature reconstruction through corresponding reverse processes, such as demultiplexing, decompression and decryption, and finally restores the original image information, thereby improving the transmission efficiency of the overall image signal.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, specifically to an adaptive encryption compression and power multiplexing transmission system based on image semantic features. Background Technology

[0002] With the development of advanced technologies such as AI, IoT, and empathic intelligence, the demand for ultra-high-speed bandwidth communication continues to grow. Multimedia data, including audio, images, and video, while enriching daily information exchange, also implies higher communication data rates and greater information security. Although many advanced communication technologies have been researched, the transmission capacity of fiber optic channels is limited by the Shannon limit theory, making breakthroughs difficult. Furthermore, facing the data deluge, personal and national data are at risk of malicious theft, making information transmission security an issue that cannot be ignored.

[0003] Existing signal transmission methods suffer from low overall transmission efficiency and security performance because the compression, encryption, and modulation stages of images are independent and they fail to differentiate processing based on the semantic importance of images. Summary of the Invention

[0004] To address the shortcomings mentioned in the background section, the present invention aims to provide an adaptive encryption compression and power multiplexing transmission system based on image semantic features.

[0005] Firstly, the objective of this invention can be achieved through the following technical solution: an image semantic feature-based adaptive encryption compression and power multiplexing transmission system, comprising: The saliency detection module is used to perform saliency detection on the original image and obtain the saliency results of different image features; The image feature classification module is used to classify the saliency results of different image features to obtain the image semantic feature classification results. An adaptive encryption and compression module is used to compress and encode the image semantic feature classification results and perform chaotic encryption to obtain encrypted and compressed image data. The power division multiplexing modulation module is used to allocate encrypted and compressed image data based on the hierarchical results of image semantic features, and to perform power domain multiplexing modulation on channels with different power levels to realize the transmission of image signals.

[0006] In conjunction with the first aspect, in some implementations of the first aspect, the system further includes: the processing procedure of the saliency detection module, as follows: The Euclidean distance between each pixel of the original image and the average feature vector of the image in the LAB color space model is calculated using the Frequency Tuned algorithm as the saliency result of different image features, as follows: in The average feature vector of the image. Let be the pixel vector at coordinates (x, y). The L2 norm is used as the Euclidean distance. In the LAB color space model, each pixel is represented as a three-dimensional vector of (L, a, b), where L represents luminance, a represents the red-green component, and b represents the blue-yellow component.

[0007] In conjunction with the first aspect, in some implementations of the first aspect, the system further includes: the image grading process of the image feature grading module includes: Based on the calculated saliency results of different image features, the original image is divided into multiple image patches, each containing... The image is classified into regions of 100 pixels, and the image grading is determined by calculating the average saliency of all pixels in each region. The calculation formula is as follows: Importance levels are classified based on the average saliency value of image patches: assuming there are three thresholds. T 1 , T 2 , T 3 ,when S mn > T 3 When: the image features of this region are considered the most important; when T 3 > S mn > T 2 When: the image features of this region are considered the second most important; when T 2 > S mn > T 1 When: the image features of this region are considered the third most important; when T 1 > S mn At that time, the image features of this region were considered the fourth most important.

[0008] In conjunction with the first aspect, in some implementations of the first aspect, the system further includes: the processing procedure of the adaptive encryption compression module, including: The image feature signals are linearly transformed through a compressed sampling matrix to obtain compressed image feature information, which is then converted into a binary representation. When the image feature is the fourth most important, the binary-represented image feature information is XORed with the chaotic sequence to obtain chaotic encrypted image data. When the image features are the second or third most important, the binary-represented image feature information and the chaotic sequence are subjected to a two-layer chaotic encryption operation to obtain the two-layer chaotic encrypted image data. When image features are of paramount importance, the binary-represented image feature information is combined with the chaotic sequence to perform a three-layer chaotic encryption operation, resulting in three-layer chaotic encrypted image data. The chaotic encrypted image data, the two-layer chaotic encrypted image data, and the three-layer chaotic encrypted image data are used together as the encrypted and compressed image data output by the adaptive encryption and compression module.

[0009] In conjunction with the first aspect, in some implementations of the first aspect, the system further includes: the calculation process of the adaptive encryption compression module, comprising: The signal x in the characteristic region undergoes a linear transformation through the compressed sampling matrix Φ: Where the matrix Also known as the measurement matrix, it is generated using the Hadamard matrix; the signal before compression. compressed signal The compression ratio is ; First, the compressed image feature information is converted into binary. This means that when the image features are at the fourth level of importance, a layer of chaotic encryption is performed. Then, driven by the key, the Lorenz chaotic model generates a chaotic sequence of the corresponding length. to Perform bitwise XOR operation: For the second most important image feature, the compressed image feature information is converted into binary... After representation, a Lorenz chaotic sequence of the corresponding length is truncated and subjected to a two-layer chaotic encryption operation: in It is the signal vector after bitwise XOR operation. This is the second set of chaotic sequences after processing. This indicates that the elements in matrix i are randomly sorted based on the values ​​in matrix j. For the most important image feature, A value of 1 is used to perform three layers of chaotic encryption on the binary representation of image features to ensure that critical information cannot be illegally stolen. Among them, the image pixels after the first two layers of encryption The corresponding coordinates are (x, y). Then, the pixel coordinates of the feature region of the image are encrypted and perturbed, and the transformed pixel coordinates are (x', y'). e and l are two independent chaotic sequences obtained by iteratively obtaining the chaotic model.

[0010] In conjunction with the first aspect, in some implementations of the first aspect, the system further includes: the data allocation process of the power division multiplexing modulation module, as follows: The first and second important image features are modulated into QAM signals and allocated to a high-power channel; The third and fourth important image features are modulated into QAM signals and assigned to low-power channels.

[0011] In conjunction with the first aspect, in some implementations of the first aspect, the system further includes: the transmitted signal after superposition coding multiplexing by the power division multiplexing modulation module is represented as: Where p1 and p2 represent the allocated signal power, and p1 > p2; s, s1, and s2 represent the multiplexed signal, the high-power signal, and the low-power signal, respectively.

[0012] In conjunction with the first aspect, in some implementations of the first aspect, the system further includes: a receiver module for demultiplexing modulation, decompression encoding, and key recovery of the signal, ultimately achieving the reconstruction and restoration of the original image.

[0013] In conjunction with the first aspect, in some implementations of the first aspect, the system further includes: the processing procedure of the receiver module, including: After transmission through the fiber optic channel, the received signal is represented as: in h Indicates the channel response. n The fiber optic channel is subject to Gaussian white noise. During SiC decoding, the signal from the high-power channel is recovered first. s 1 The low-power channel signal at this time is treated as noise; then, the received signal is extracted... s1 Signal subtraction restores low-power signal s 2； After demultiplexing modulation, the signal needs further decompression and encoding. This is because the encryption and compression process at the transmitting end requires a measurement matrix that satisfies the constrained isochronous property. There exists a constant satisfy: The receiver then utilizes the sparsity of the signal to reconstruct the original signal by minimizing the regularized objective function: in It is a regularization parameter. y It is the observed signal vector. It is the signal vector to be recovered.

[0014] The beneficial effects of this invention are: This invention effectively reduces redundant information in the data source while protecting important semantic features through adaptive encryption compression based on image semantic features; it also performs layered chaotic encryption on different image features to ensure the security of critical data information. Considering limited spectrum resources, it constructs differentiated power multiplexing modulation for image feature regions, prioritizing the protection of information in key image feature regions and improving the overall signal transmission efficiency. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 This is a schematic diagram of the processing flow of the present invention; Figure 3 This is a schematic diagram of the significance detection results of the present invention; Figure 4 This is a schematic diagram of the image feature grading results of the present invention; Figure 5 This is a schematic diagram of the image after adaptive encryption and compression according to the present invention; Figure 6 This is a schematic diagram of the image information recovered by the receiving end of the present invention; Figure 7 This is a schematic diagram of the bit error rate curve of the receiver of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Example 1: like Figure 1 As shown, an image semantic feature-based adaptive encryption compression and power multiplexing transmission system includes: The saliency detection module is used to perform saliency detection on the original image and obtain the saliency results of different image features; like Figure 3 As shown, in the saliency detection module, the Frequency Tuned algorithm can be used to calculate the Euclidean distance between each pixel of the original image in the LAB color space model and the average feature vector of the image: in The average feature vector of the image. Let be the pixel vector at coordinates (x, y). The L2 norm (i.e., Euclidean distance) is used. In the LAB color space model, each pixel can be represented as a three-dimensional vector of (L, a, b), where L represents luminance, a represents the red-green component, and b represents the blue-yellow component.

[0018] The saliency detection module ultimately calculates the overall differences in the image and effectively detects key feature regions in the image.

[0019] The image feature classification module is used to classify the saliency results of different image features to obtain the image semantic feature classification results. The image grading process of the image feature grading module includes: like Figure 4 As shown, based on the calculated saliency results of different image features, the original image is divided into multiple image patches, each containing... The image is classified into regions of 100 pixels, and the image grading is determined by calculating the average saliency of all pixels in each region. The calculation formula is as follows: Importance levels are classified based on the average saliency value of image patches: assuming there are three thresholds. T 1 , T 2 , T3 ,when S mn > T 3 When: the image features of this region are considered the most important; when T 3 > S mn > T 2 When: the image features of this region are considered the second most important; when T 2 > S mn > T 1 When: the image features of this region are considered the third most important; when T 1 > S mn At that time, the image features of this region were considered the fourth most important.

[0020] Among them, three thresholds T 1 , T 2 , T 3 The settings are mainly based on the actual image saliency results and the image feature regions of most interest. S at Background area S ba Make dynamic adjustments: a. Initialization T 1 , T 2 , T 3 Numerical value; b. Comparison S at and T 3 ,when S at < T 3 ,renew T 3 = S at Otherwise, keep T 3 Ensure the importance level of the most relevant image feature regions; c. Comparison S ba and T 1 ,when S ba > T1 ,renew T 1 = S ba Otherwise, keep T 1 Ensure that the background area of ​​the image has a low importance level.

[0021] Among them, the first most important is the most important, the second most important is the second most important, the third most important is the moderately important, and the fourth most important is the least important. The degree of importance decreases as the number increases. An adaptive encryption and compression module is used to compress and encode the image semantic feature classification results and perform chaotic encryption to obtain encrypted and compressed image data. like Figure 5 As shown, the processing procedure of the adaptive encryption compression module includes: The image feature signals are linearly transformed through a compressed sampling matrix to obtain compressed image feature information, which is then converted into a binary representation. When the image feature is the fourth most important, the binary-represented image feature information is XORed with the chaotic sequence to obtain chaotic encrypted image data. When the image features are the second or third most important, the binary-represented image feature information and the chaotic sequence are subjected to a two-layer chaotic encryption operation to obtain the two-layer chaotic encrypted image data. When image features are of paramount importance, the binary-represented image feature information is combined with the chaotic sequence to perform a three-layer chaotic encryption operation, resulting in three-layer chaotic encrypted image data. The chaotic encrypted image data, the two-layer chaotic encrypted image data, and the three-layer chaotic encrypted image data are used together as the encrypted and compressed image data output by the adaptive encryption and compression module.

[0022] The signal x in the characteristic region undergoes a linear transformation through the compressed sampling matrix Φ: Where the matrix Also known as the measurement matrix, it is generated using the Hadamard matrix; the signal before compression. compressed signal The compression ratio is ; First, the compressed image feature information is converted into binary. This means that when the image features are at the fourth level of importance, a layer of chaotic encryption is performed. Then, driven by the key, the Lorenz chaotic model generates a chaotic sequence of the corresponding length. to Perform bitwise XOR operation: For the second most important image feature The value is relatively high to avoid the loss of key image features due to compression. The compressed image feature information is then converted into binary... After representation, a Lorenz chaotic sequence of the corresponding length is truncated and subjected to a two-layer chaotic encryption operation: in It is the signal vector after bitwise XOR operation. It is the second set of chaotic sequences after processing (its value range is positive integers [1~M]). This indicates that the elements in matrix i are randomly sorted based on the values ​​in matrix j. For the most important image feature, A value of 1 ensures optimal transmission of critical image semantic features. Furthermore, three layers of chaotic encryption are applied to the binary representation of image features to guarantee that critical information cannot be illegally intercepted. Among them, the image pixels after the first two layers of encryption The corresponding coordinates are (x, y). Then, the pixel coordinates of the feature region of the image are encrypted and perturbed, and the transformed pixel coordinates are (x', y'). e and l are two independent chaotic sequences obtained by iteratively obtaining the chaotic model.

[0023] The adaptive encryption compression module implements differentiated compression encoding of image semantic features, which effectively preserves the semantic information of high-level image features and the security of transmission, while reducing the amount of transmitted data for image blocks containing less semantic information.

[0024] The power division multiplexing modulation module is used to allocate encrypted and compressed image data based on the hierarchical results of image semantic features, and to perform power domain multiplexing modulation on channels with different power levels to realize the transmission of image signals.

[0025] Compared to traditional orthogonal multiple access (OMA) techniques, non-orthogonal multiple access (NOR) offers better spectral efficiency, and power division multiplexing (PDM), as a typical NOM technique, exhibits excellent compatibility with existing multicarrier modulation schemes. PDM distinguishes signals by allocating different powers, and the receiver utilizes a serial interference cancellation (SIC) algorithm to detect the demultiplexed signal.

[0026] In this application, the receiving end needs to select a demodulation algorithm for power division multiplexing modulation, and finally selects the serial interference cancellation (SIC) algorithm. It prioritizes demodulating the high-power signal. After demodulating the high-power signal, it is subtracted from the overall signal, leaving the low-power signal. Then, the low-power signal is modulated to recover the low-power signal, thereby enabling the demodulation of the power division multiplexing modulation signal.

[0027] In this invention patent, the power division multiplexing modulation module considers the semantic feature level of the image signal to be modulated, modulates the "most important" and "second most important" image semantic signals into QAM signals and allocates them to the high-power channel; the QAM signals obtained by modulating the "generally important" and "unimportant" image semantic signals are allocated to the low-power channel. Finally, the transmitted signal after superposition coding multiplexing can be expressed as: Where p1 and p2 represent the allocated signal power, and p1 > p2; in actual communication scenarios, p1 and p2 need to be reasonably allocated according to channel conditions to meet the requirements of the current communication scenario. s, s1, and s2 represent the multiplexed signal, the high-power signal, and the low-power signal, respectively.

[0028] It also includes a receiver module, which is used to demultiplex the signal, decompress and encode it, and recover the key, ultimately realizing the reconstruction and restoration of the original image.

[0029] Specifically, after transmission through the optical fiber channel, the received signal can be represented as: Where h represents the channel response and n is the Gaussian white noise of the fiber optic channel. During SIC decoding, the high-power channel signal, i.e., s1, is recovered first, while the low-power channel signal is treated as noise. Then, s1 is subtracted from the received signal to recover the low-power signal s2.

[0030] After demultiplexing modulation, the signal needs further decompression and encoding. This is because the encryption and compression process at the transmitting end requires a measurement matrix that satisfies the constrained isochronous property. There exists a constant satisfy: The receiver can then leverage the sparsity of the signal to reconstruct the original signal by minimizing the regularization objective function. in Here, y is the regularization parameter, and y is the observed signal vector. This is the signal vector to be recovered. Through iterative optimization, the SL0 algorithm can implement the corresponding decompression and encoding process at the receiving end.

[0031] Specifically, the following embodiments further illustrate the solution of the present invention: The final decompressed and encoded signal vector undergoes key recovery and image feature reconstruction to obtain the original image information. For example... Figure 6 The image shown is the image information recovered by this invention after transmission in a Gaussian channel with a signal-to-noise ratio of 22 dB using MATLAB simulation software. Compared with the original image, the key information in the image is preserved, while less important image parts show varying degrees of blurring. Figure 6 (b) shows the recovered image information after the receiving end loses some key information. It can be found that the key semantic features of the image cannot be correctly decrypted to extract the information content. This demonstrates that the scheme provides good security protection for high-priority image semantic information.

[0032] Figure 7 The paper further illustrates the bit error rate (BER) curves and multi-scale structural similarity (MSSIM) curves at the receiver under different signal-to-noise ratio (SNR) channel conditions. As the SNR increases, the BER at the receiver gradually decreases, while the multi-scale structural similarity between the recovered image and the original image increases. Notably, the BER performance of high-power signals is superior to that of low-power signals. Therefore, modulating higher-priority image feature information onto high-power signals for transmission contributes to the complete transmission of image semantic features.

[0033] Simulation results show that in the adaptive encryption compression-power multiplexing signal transmission system based on image semantic features, combining image semantic features with encryption compression and power multiplexing modulation achieves adaptive encryption compression and differentiated multiplexing modulation of image features, retains rich salient image semantic feature information, and realizes secure and efficient transmission of salient image semantic information, which can take into account both the important semantic transmission of the image and the overall signal transmission efficiency.

[0034] Based on the same inventive concept, this invention also provides a computer device, comprising: one or more processors, and a memory for storing one or more computer programs; the programs include program instructions, and the processor executes the program instructions stored in the memory. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, used to implement one or more instructions, specifically for loading and executing one or more instructions stored in a computer storage medium to implement the above-described method.

[0035] It should be further explained that, based on the same inventive concept, the present invention also provides a computer storage medium storing a computer program, which, when executed by a processor, performs the above-described method. This storage medium can be any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In the present invention, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0036] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this disclosure. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0037] The foregoing has shown and described the basic principles, main features, and advantages of this disclosure. Those skilled in the art should understand that this disclosure is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of this disclosure. Various changes and modifications can be made to this disclosure without departing from its spirit and scope, and all such changes and modifications fall within the scope of this disclosure as claimed.

Claims

1. An image semantic feature-based adaptive encryption compression and power multiplexing transmission system, characterized in that, include: The saliency detection module is used to perform saliency detection on the original image and obtain the saliency results of different image features; The image feature classification module is used to classify the saliency results of different image features to obtain the image semantic feature classification results. An adaptive encryption and compression module is used to compress and encode the image semantic feature classification results and perform chaotic encryption to obtain encrypted and compressed image data. The power division multiplexing modulation module is used to allocate encrypted and compressed image data based on the hierarchical results of image semantic features, and to perform power domain multiplexing modulation on channels with different power levels to realize the transmission of image signals.

2. The image semantic feature-based adaptive encryption compression and power multiplexing transmission system according to claim 1, characterized in that, The processing procedure of the saliency detection module is as follows: The Euclidean distance between each pixel of the original image and the average feature vector of the image in the LAB color space model is calculated using the Frequency Tuned algorithm as the saliency result of different image features, as follows: in The average feature vector of the image. Let be the pixel vector at coordinates (x, y). The L2 norm is used as the Euclidean distance. In the LAB color space model, each pixel is represented as a three-dimensional vector of (L, a, b), where L represents luminance, a represents the red-green component, and b represents the blue-yellow component.

3. The image semantic feature-based adaptive encryption compression and power multiplexing transmission system according to claim 1, characterized in that, The image grading process of the image feature grading module includes: Based on the calculated saliency results of different image features, the original image is divided into multiple image patches, each containing... The image is classified into regions of 100 pixels, and the image grading is determined by calculating the average saliency of all pixels in each region. The calculation formula is as follows: Importance levels are classified based on the average saliency value of image patches: assuming there are three thresholds. T 1 , T 2 , T 3 ,when S mn > T 3 When: the image features of this region are considered the most important; when T 3 > S mn > T 2 When: the image features of this region are considered the second most important; when T 2 > S mn > T 1 When: the image features of this region are considered the third most important; when T 1 > S mn At that time, the image features of this region were considered the fourth most important.

4. The image semantic feature-based adaptive encryption compression and power multiplexing transmission system according to claim 1, characterized in that, The processing procedure of the adaptive encryption compression module includes: The image feature signals are linearly transformed through a compressed sampling matrix to obtain compressed image feature information, which is then converted into a binary representation. When the image feature is the fourth most important, the binary-represented image feature information is XORed with the chaotic sequence to obtain chaotic encrypted image data. When the image features are the second or third most important, the binary-represented image feature information and the chaotic sequence are subjected to a two-layer chaotic encryption operation to obtain the two-layer chaotic encrypted image data. When image features are of paramount importance, the binary-represented image feature information is combined with the chaotic sequence to perform a three-layer chaotic encryption operation, resulting in three-layer chaotic encrypted image data. The chaotic encrypted image data, the two-layer chaotic encrypted image data, and the three-layer chaotic encrypted image data are used together as the encrypted and compressed image data output by the adaptive encryption and compression module.

5. The image semantic feature-based adaptive encryption compression and power multiplexing transmission system according to claim 4, characterized in that, The calculation process of the adaptive encryption compression module includes: The signal x in the characteristic region undergoes a linear transformation through the compressed sampling matrix Φ: Where the matrix Also known as the measurement matrix, it is generated using the Hadamard matrix; the signal before compression. compressed signal The compression ratio is ; First, the compressed image feature information is converted into binary. This means that when the image features are at the fourth level of importance, a layer of chaotic encryption is performed. Then, driven by the key, the Lorenz chaotic model generates a chaotic sequence of the corresponding length. , and Perform bitwise XOR operation: For the second most important image feature, the compressed image feature information is converted into binary... After representation, a Lorenz chaotic sequence of the corresponding length is truncated and subjected to a two-layer chaotic encryption operation: in It is the signal vector after bitwise XOR operation. This is the second set of chaotic sequences after processing. This indicates that the elements in matrix i are randomly sorted based on the values ​​in matrix j. For the most important image feature, A value of 1 is used to perform three layers of chaotic encryption on the binary representation of image features to ensure that critical information cannot be illegally stolen. Among them, the image pixels after the first two layers of encryption The corresponding coordinates are (x, y). Then, the pixel coordinates of the feature region of the image are encrypted and perturbed, and the transformed pixel coordinates are (x', y'). e and l are two independent chaotic sequences obtained by iteratively obtaining the chaotic model.

6. The image semantic feature-based adaptive encryption compression and power multiplexing transmission system according to claim 1, characterized in that, The data allocation process of the power division multiplexing modulation module is as follows: The first and second important image features are modulated into QAM signals and allocated to a high-power channel; The third and fourth important image features are modulated into QAM signals and assigned to low-power channels.

7. The image semantic feature-based adaptive encryption compression and power multiplexing transmission system according to claim 6, characterized in that, The transmitted signal after superposition coding multiplexing by the power division multiplexing modulation module is represented as follows: Where p1 and p2 represent the allocated signal power, and p1 > p2; s, s1, and s2 represent the multiplexed signal, the high-power signal, and the low-power signal, respectively.

8. The image semantic feature-based adaptive encryption compression and power multiplexing transmission system according to claim 1, characterized in that, It also includes a receiver module, which is used to demultiplex the signal, decompress and encode it, and recover the key, ultimately realizing the reconstruction and restoration of the original image.

9. The image semantic feature-based adaptive encryption compression and power multiplexing transmission system according to claim 8, characterized in that, The processing procedure of the receiver module includes: After transmission through the fiber optic channel, the received signal is represented as: in h Indicates the channel response. n The fiber optic channel is subject to Gaussian white noise. During SiC decoding, the signal from the high-power channel is recovered first. s 1 The low-power channel signal at this time is treated as noise; then, the received signal is extracted... s 1 Signal subtraction restores low-power signal s 2； After demultiplexing modulation, the signal needs further decompression and encoding. This is because the encryption and compression process at the transmitting end requires a measurement matrix that satisfies the constrained isochronous property. There exists a constant satisfy: The receiver then utilizes the sparsity of the signal to reconstruct the original signal by minimizing the regularization objective function: in It is a regularization parameter. y It is the observed signal vector. It is the signal vector to be recovered.

10. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that, The memory stores a computer program that can run on the processor. When the processor loads and executes the computer program, it employs the image semantic feature-based adaptive encryption compression and power multiplexing transmission system as described in any one of claims 1 to 8.

Images