Image transmission system and method applied to flight attitude control and encryption

Abstract

The invention belongs to the technical field of flight attitude control and encrypted image transmission, and provides an image transmission system and method applied to flight attitude control and encryption, and the method comprises the following steps: grouping data, distributing an encryption key for each group of data, and then carrying out secondary encryption operation by means of an elliptic curve encryption technology. Through a dynamic modulation strategy based on reinforcement learning driving, historical ionospheric channel data are mined through a deep Q network, a Shannon capacity model is introduced to verify reinforcement learning decisions, when high-order modulation exceeds a channel theoretical limit, automatic order reduction is performed, the error rate exceeding risk is reduced, dynamic balance of transmission reliability and efficiency is ensured, and the method is suitable for large-scale popularization and application. According to the polarization diversity technology, horizontal and vertical polarization signals are emitted at the same time, the probability of simultaneous deep fading of the signals is reduced from traditional single-polarization transmission by utilizing the independent fading characteristic of an ionized layer on different polarization signals, and the stability of communication between the unmanned aerial vehicle and a satellite is improved by combining dynamic power distribution and high-precision pilot frequency synchronization.

Description

Technical Field

[0001] The present invention belongs to the technical field of flight attitude control and encrypted image transmission, and specifically relates to an image transmission system and method applied to flight attitude control and encryption. Background Art

[0002] In today's era of rapid technological development, flight attitude control and encrypted image transmission technology play a vital role in many fields, such as national defense and military, aerospace exploration, and civil drones. In terms of data security encryption, traditional encryption methods mostly use AES-256 static encryption algorithm, which lacks flexibility in key management. Once the key is leaked, the entire encryption system will face the risk of collapse. With the development of quantum computing technology, its encryption strength has been unable to resist the potential threat of quantum computing attacks. For flight attitude control and encrypted image transmission systems, data confidentiality and integrity are crucial; Ionospheric communication is one of the ways to transmit data in flight, but the characteristics of the ionosphere pose a great challenge to communication. Parameters such as the electron density and temperature of the ionosphere are in a highly dynamic state, which makes the quality of the communication channel extremely unstable. The traditional fixed modulation method cannot be adaptively adjusted according to the changes in the ionosphere. When the channel conditions are poor, data transmission is prone to bit errors and packet loss, affecting the communication quality. For example, during the peak of solar activity, the drastic changes in the ionosphere will lead to increased signal fading. If a communication system using a fixed modulation method is used, large-scale communication interruptions may occur, and stable communication between drones and command centers, satellites and ground stations cannot be guaranteed. In addition, the high-speed movement of satellites produces the Doppler frequency shift effect, which will cause the frequency of the received signal to shift, making data demodulation difficult. If the frequency shift cannot be accurately compensated, the data received by the receiving end will be erroneous, affecting the accurate acquisition of information. In addition, satellite signals are susceptible to malicious interference and hijacking. Traditional communication technologies have limited means to deal with these threats. In military confrontation scenarios, the enemy may interfere with satellite signals and destroy the communication link between the drone and the command center, causing the drone to lose control, thereby affecting the smooth progress of combat operations. To this end, the present invention provides an image transmission system and method for flight attitude control and encryption. Summary of the invention

[0003] In order to make up for the deficiencies of the prior art, at least one technical problem raised in the background technology is solved.

[0004] The technical solution adopted by the present invention to solve the technical problem is: an image transmission method applied to flight attitude control and encryption, comprising the following steps: Group the data, assign an encryption key to each group of data, and then perform a secondary encryption operation using elliptic curve encryption technology. At the same time, use data redundancy coding technology to embed a specific proportion of redundant information in the encrypted data; After the encryption process is completed, monitor the channel conditions of the ionosphere in real time, dynamically adjust the modulation mode of the data, and use polarization diversity technology to transmit horizontal polarization and vertical polarization signals simultaneously; Monitor the frequency of the satellite signal in real time, calculate the total Doppler frequency shift, adjust the frequency of the received signal according to the total Doppler frequency shift, and deploy frequency hopping spread spectrum technology. The transmitter quickly hops and transmits data on different frequency channels according to a predetermined pseudo-random sequence; Use the predicted attitude map comparison mechanism and the motion trajectory prediction model to reverse-infer the complete attitude data and generate control commands, and divide the priorities of the control commands, so that the key control data with high priority skips the clustering process and directly enters the encryption channel.

[0005] An image transmission system applied to flight attitude control and encryption specifically includes: Multi-level encryption processing module: Group the data, assign an encryption key to each group of data, and then perform a secondary encryption operation using elliptic curve encryption technology. At the same time, use data redundancy coding technology to embed a specific proportion of redundant information in the encrypted data; Ionosphere channel processing module: After the encryption process is completed, monitor the channel conditions of the ionosphere in real time, dynamically adjust the modulation mode of the data, and use polarization diversity technology to transmit horizontal polarization and vertical polarization signals simultaneously; Satellite signal processing module: Monitor the frequency of the satellite signal in real time, calculate the total Doppler frequency shift, adjust the frequency of the received signal according to the total Doppler frequency shift, and deploy frequency hopping spread spectrum technology. The transmitter quickly hops and transmits data on different frequency channels according to a predetermined pseudo-random sequence; Instruction priority processing module: Use the predicted attitude map comparison mechanism and the motion trajectory prediction model to reverse-infer the complete attitude data and generate control commands, and divide the priorities of the control commands, so that the key control data with high priority skips the clustering process and directly enters the encryption channel.

[0006] The beneficial effects of the present invention are as follows: 1. The present invention optimizes the problem of inefficient grouping of traditional fixed clustering methods when data changes non-steadily by converting the data to be transmitted into a data graph, dynamically clustering based on geometric and statistical features, adaptively generating and efficiently grouping encryption keys, automatically switching clustering algorithms according to the feature dispersion of stationary and complex maneuvering data, reducing manual intervention and computational overhead. The elliptic curve encryption technology is used for secondary encryption, significantly enhancing the security of the key, achieving a balance between security and efficiency on low-computing-power terminals such as drones. The redundant coding technology embeds check codes by combining the key vertices of the contour line and the time-series difference of eigenvalues. Compared with full-data redundancy, it reduces bandwidth overhead while ensuring the error correction ability. 2. The present invention adopts a dynamic modulation strategy driven by reinforcement learning, mines historical ionospheric channel data through a deep Q-network, introduces the Shannon capacity model to verify the reinforcement learning decision, automatically reduces the order when high-order modulation exceeds the channel theoretical limit, reduces the risk of excessive bit error rate, and ensures the dynamic balance between transmission reliability and efficiency. The polarization diversity technology transmits horizontal and vertical polarization signals simultaneously, and uses the independent fading characteristics of the ionosphere for different polarization signals to reduce the probability of simultaneous deep fading of signals from that of traditional single-polarization transmission. Combined with dynamic power allocation and high-precision pilot synchronization, it improves the stability of drone-satellite communication. 3. The present invention optimizes the calculation of the total Doppler frequency shift, optimizes the problem of signal demodulation deviation caused by high-speed movement, adopts a chaotic frequency-hopping sequence and dynamic frequency-hopping parameter adjustment to enhance the anti-predictive interference ability of the system. The dual-frequency pilots inserted in the frequency-hopping OFDM signal provide multi-dimensional synchronization information for the receiving end. When the ionospheric link quality deteriorates, it automatically increases the frequency-hopping power of satellite communication and reduces the interval to achieve seamless switching between the ionosphere and satellite communication. 4. The present invention constructs a prediction model and a dynamic time warping algorithm based on historical attitude data to divide the real-time priority of control commands. For key control data with a similarity higher than the high-accuracy limit value, it can skip the clustering process and directly enter the encryption channel. This dynamic priority mechanism optimizes the limitations of traditional fixed-priority schemes. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] The present invention will be further described below with reference to the accompanying drawings.

[0008] Figure 1 is the flowchart of the steps of the method for transmitting images with flight attitude control and encryption of the present invention; Figure 2 is the framework diagram of the system for transmitting images with flight attitude control and encryption of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0009] In order to make the technical means, creative features, achieved purposes and effects of the present invention easy to understand, the present invention will be further described below in conjunction with specific embodiments.

[0010] Example 1: Please refer to Figure 1 As shown, the embodiment of the present invention in the flight attitude control and encrypted video transmission method includes the following steps: Step 1: Apply a multi-level asymmetric encryption algorithm system to group the data, assign an encryption key to each group of data, then perform a secondary encryption operation using elliptic curve encryption technology, and at the same time, use data redundancy coding technology to embed a specific proportion of redundant information in the encrypted data; In this step, first, pack the data to be transmitted according to preset data conditions; the data conditions are related to the minimum configuration and communication requirements of the requesting end; convert the data to be transmitted into a data graph, compare the data graphs, and determine a preset number of node graphs; Converting the data to be transmitted can obtain relatively simple data that is convenient for analysis, that is, the data graph. Comparing the data graphs corresponding to multiple data to be transmitted within the same packing result, a representative data graph can be selected as the reference node, called the node graph; Specifically, the steps of converting the data to be transmitted into a data graph, comparing the data graphs, and determining a preset number of node graphs include: S101: Calculate the eigenvalue of each data to be transmitted according to a preset fitting formula to obtain an eigenvalue group in time order; S102: Traverse the data to be transmitted, and perform binary processing on each pixel point according to preset color value conditions to obtain a data graph; S103: Perform pixel clustering on the data graph to determine the contour lines of each type of pixel point; S104: Compare the data graphs containing contour lines, and determine a preset number of node graphs according to the comparison result and the eigenvalue group; Through the above S101 - S103, the original data to be transmitted is converted into a structured data graph (including eigenvalue sequence and pixel contour features), providing quantifiable feature dimensions for grouping (such as time-order features, pixel distribution features). Through S104, representative reference data units are selected from the data graph as the core basis for grouping; for example, by comparing the contour lines and eigenvalue groups, similar data is grouped into the same group to ensure the consistency of each group of data in the feature space, which is convenient for subsequent unified distribution of encryption keys; The process of assigning an encryption key to each group of data is as follows: Each group of data corresponds to one or more node graphs. Perform a joint hash (such as SHA - 256) on the eigenvalue group (time-order feature) and the contour line coordinate sequence of the node graph to generate a hash value of a fixed length; Using the hash value as a seed, generate a symmetric encryption key K_group through a key derivation function (KDF, such as PBKDF2) for the initial encryption of this group of data; Adjust the key length according to the minimum configuration of the requesting end (such as computing power, memory limit), including 128 bits and 256 bits, to ensure a balance between encryption efficiency and security; Each group of data corresponds to a unique grouping key K_group for subsequent data encryption; The process of performing secondary encryption using elliptic curve encryption technology is as follows: Select standard elliptic curve parameters (such as the P-256 curve recommended by NIST) to generate the public and private key pairs of the receiving party; Furthermore, reduce the direct transmission of the grouping key K_group and perform secondary encapsulation on K_group through elliptic curve encryption technology; Furthermore, the process of secondary encryption is specifically as follows: The sender encrypts the grouping key K_group with the elliptic curve public key EC_pub of the receiving party to obtain the ciphertext, and binds the original data (encrypted with the grouping key) to C_EC to form a secondary encrypted data unit; Through the previous process of secondary encryption, even if the grouping key is intercepted during transmission, the attacker needs to solve the elliptic curve discrete logarithm problem to obtain K_group, significantly improving the security of key transmission; The process of embedding a specific proportion of redundant information in the encrypted data is as follows: Determine the redundancy ratio according to communication requirements (such as network bit error rate, real-time requirements); If the communication link stability of the requesting end is poor, increase the redundancy ratio; if sensitive to bandwidth, reduce the redundancy ratio, and balance the redundancy overhead through the complexity of the node graph features (such as the number of vertices of the contour line); Furthermore, the process of generating and embedding redundant information is as follows: Perform error correction coding (such as Reed-Solomon code or low-density parity-check code LDPC) on the coordinate sequence of the contour line of the data graph to generate a redundant check sequence; Embed the redundant check sequence into the encrypted data according to a preset rule; The prediction rule can specifically be: Insert check bytes into the ciphertext area corresponding to the key vertices of the contour line; Insert differential check codes between adjacent eigenvalues for the time-series data in the eigenvalue group; Key vertices include but are not limited to: inflection points, extreme points; Record the redundancy embedding rule in the metadata, and the receiving end can quickly locate the redundant area through the node graph features for efficient error correction; As another implementation manner of this embodiment: Generate a data graph containing contour lines, including the binarization result of pixel points and the coordinate sequence of the contour line; S201: Extract geometric features and statistical features to form a multi-dimensional feature vector. Geometric features include: Calculate the contour perimeter (total pixel side length), contour area (number of pixels in the closed area), eccentricity (degree to which the contour shape approaches an ellipse), rectangularity (ratio of the contour area to the area of the minimum circumscribed rectangle); The statistical features include: the mean, variance, and extreme values of the extracted contour vertex coordinates; Each contour line data graph corresponds to a comprehensive feature vector; S202: Dynamically switch the clustering strategy according to the characteristics of the flight data. If it is stable flight data, use the DBSCAN algorithm with high computational efficiency; if it is complex maneuvering data, use the self-organizing map SOM; Specifically: The process of determining the characteristics of the flight data is: calculating the dispersion of the feature vector distribution; For the feature vector set of the current batch of real-time data, the feature vector set includes: geometric features and statistical features; Normalize the feature vectors of different dimensions in the feature vector set; Calculate the global dispersion index, and the dispersion index is the coefficient of variation. Among them, the coefficient of variation represents the relative dispersion degree of the data in this dimension, and the larger the value, the more dispersed the data distribution; Integrate the dispersion characteristics of all dimensions, and use the weighted average coefficient of variation as the global dispersion index; Those skilled in the art set the weight coefficients for the feature vectors of each dimension according to the sensitivity of the flight state, and perform weighted fusion calculation, and the obtained value is used as the global dispersion index; Set the stable threshold and the complex threshold respectively. If the global dispersion index is less than or equal to the stable threshold, it means that the distribution is concentrated, then it is determined as stable flight data, and the DBSCAN algorithm is enabled; if the global dispersion index is greater than or equal to the complex threshold, it means that the data distribution is dispersed, then it is determined as complex maneuvering data, and the self-organizing map SOM algorithm is enabled; if the global dispersion index is greater than the stable threshold and less than the complex threshold, use the clustering strategy of the previous batch; Optionally, to reduce misjudgment caused by a single instance, the global dispersion index of consecutive N batches can be calculated and averaged to obtain the global dispersion index mean, and the determination process of the flight characteristics is based on the global dispersion index mean; Input the feature vector into the clustering model, preset the range of the initial number of clusters, and automatically evaluate the optimal number of clusters through the Silhouette Score or the Davies-Bouldin index; S203: Dynamically group based on the real-time flight data distribution. During the flight, every time a batch of data graphs is received, perform the following process: Extract the comprehensive feature vector from the new data graph and merge it with the historical feature library to form a new comprehensive feature vector; The clustering model performs real-time clustering on F_set_new and outputs the current optimal number of node graphs; Select the "centroid data graph" of each group as the node graph according to the clustering result; The dynamically generated set of node graphs, as well as the feature mean and distribution range of each group of data; In summary, through the collaborative design of data preprocessing, dynamic grouping, encryption, and redundant coding, this step realizes secure and efficient data transmission; First, the data to be transmitted is packaged according to the configuration of the requesting end, and a data graph with contour lines is generated through eigenvalue calculation (S101), binarization processing (S102), and MeanShift clustering (S103); then, geometric and statistical features are extracted to form a multi-dimensional vector (S201), the data characteristics are determined by calculating the weighted mean of the coefficient of variation, and the DBSCAN or SOM clustering algorithm is dynamically switched (S202), and the optimal number of node graphs and the reference node graph are adaptively determined (S203); a grouping key is generated based on the characteristics of the node graph and adapted to the configuration of the requesting end, the key is secondarily encapsulated through elliptic curve encryption technology, and finally, in combination with the complexity of the node graph and the communication requirements, targeted redundant check codes are embedded in the key areas of the contour line and the eigenvalue time series to form a complete closed loop of "data grouping - encryption - error correction"; It has the following effects: by quantifying the data distribution dispersion through the coefficient of variation, automatically switching to the efficient DBSCAN algorithm or SOM algorithm, it optimizes the problem of inefficient grouping of the fixed preset clustering method when the data changes non-steadily, and reduces the manual parameter adjustment cost; The grouping key is dynamically generated based on the characteristics of the node graph and the key length is dynamically adjusted according to the computing power of the requesting end, achieving a balance between security and efficiency on low-computing-power devices such as drones; the redundant coding embeds check codes by combining the key vertices of the contour line and the eigenvalue time series difference. Compared with full-data redundancy, it reduces the bandwidth overhead while ensuring the error correction ability; The elliptic curve secondary encryption technology improves the security of key transmission. The attacker needs to solve the highly complex discrete logarithm problem. Combined with the binding design of the node graph metadata and the redundant rules, the receiving end can quickly locate the key area to achieve efficient error correction, meeting the strict requirements of flight control instructions for real-time performance and error resistance; Step 2: After completing the encryption process, the channel condition of the ionosphere is monitored in real time, the modulation mode of the data is dynamically adjusted, switched from the basic binary phase shift keying to the high-order quadrature amplitude modulation, and the polarization diversity technology is used to simultaneously transmit horizontally polarized and vertically polarized signals; In this step, at the sending end, the channel condition of the ionosphere is monitored in real time. According to parameters such as signal strength and bit error rate, the modulation mode of the data is dynamically adjusted from the basic binary phase shift keying (BPSK) to the high-order quadrature amplitude modulation (QAM) to ensure that the data is transmitted in the best way under different ionospheric environments. At the same time, using the polarization diversity technology, by simultaneously transmitting horizontally polarized and vertically polarized signals, the transmission stability of the signal in the ionosphere is increased, and the signal fading caused by ionospheric reflection and refraction is reduced; The specific steps are as follows: The ionospheric channel parameters are collected in real time by the sensors built into the transmitter. Among them, the channel parameters include, but are not limited to, signal strength, bit error rate, and channel fading coefficient. Specifically, the signal strength is the power value of the received reference signal; the bit error rate is the real-time bit error rate calculated through preamble verification; the channel fading coefficient is to estimate the horizontal polarization channel coefficient and the vertical polarization channel coefficient based on the phase change of the pilot signal, and construct a channel matrix , where is the horizontal polarization channel coefficient, is the vertical polarization channel coefficient; Dynamic decision-making in the modulation stage based on the collaboration of reinforcement learning drive and Shannon capacity model: Construct a reinforcement learning framework; The reinforcement learning framework includes: state space, action space, and reward function; Specifically, the state space S includes real-time channel parameters and historical modulation strategies. Among them, the historical modulation strategy is the modulation methods selected in the last three times; The action space is A = {BPSK, QPSK, 16QAM, 64QAM, 256QAM}, a total of 5 modulation methods; The calculation formula of the reward function R is: ; where , , are weight coefficients, is the transmission rate, which reflects the amount of data successfully transmitted per unit time. The larger this value, the higher the reward, is the bit error rate, which represents the proportion of bits with errors in the transmission. The larger this value, the higher the penalty, is the modulation method change indication function (takes 1 when changing, otherwise 0); Use historical ionospheric channel data to train a deep Q network (DQN). The input layer is the state vector, and the output layer is the Q value of each action; Input the real-time state into the trained deep Q network, select the modulation method with the largest Q value as the optimal modulation method, and at the same time, update the experience replay buffer through the demodulation result feedback from the receiver, and perform policy iteration regularly, such as setting policy iteration every 100 cycles; When the theoretical bit error rate corresponding to the modulation order recommended by reinforcement learning does not exceed the target bit error rate, calculate the theoretical channel capacity under the optimal modulation method at the same time. If the theoretical channel capacity is less than or equal to the Shannon extreme value, the optimal modulation method output by the deep Q network is adopted; if the theoretical channel capacity is greater than the Shannon extreme value, it means that the modulation method recommended by the current reinforcement learning exceeds the actual support capacity of the channel. At this time, a modulation method one level lower than the current modulation order needs to be selected from the action space. If the theoretical bit error rate of the newly selected modulation method still does not exceed the target bit error rate and the theoretical channel capacity is less than or equal to the Shannon extreme value, this modulation method is adopted; if the conditions are still not met after adjustment, repeat the above downscaling operation until a modulation method that meets the Shannon limit and has a qualified bit error rate is found; When the theoretical bit error rate corresponding to the modulation order recommended by reinforcement learning exceeds the target bit error rate, automatically return to the sub-optimal order of Shannon capacity calculation; The Shannon capacity calculation model is: , where C is the channel capacity, B is the channel bandwidth, and SNR represents the signal-to-noise ratio; by quantifying the relationship between the channel bandwidth and the signal-to-noise ratio and the channel capacity, it provides a theoretical basis for the selection of the modulation order; Traverse the action space, calculate the theoretical bit error rate for each modulation method, screen out the modulation methods that satisfy the theoretical bit error rate ≤ target bit error rate, construct a set of candidate modulation methods, calculate the theoretical channel capacity for each modulation method in the selected modulation method set, and sort all the theoretical channel capacities in descending order. The modulation method corresponding to the maximum theoretical channel capacity is used as the optimal modulation method. If there is only one modulation method in the candidate modulation method set, directly output this modulation method as the optimal modulation method; Based on the determined optimal modulation method, the transmitter adjusts its internal parameters to achieve precise switching of the modulation method; Exemplarily, if the decision is to switch from BPSK to 16QAM, the modulation module will reconfigure the constellation mapping rule so that each symbol carries 4 bits of information (compared with 1 bit of BPSK), greatly improving the spectral efficiency. If switching to the higher-order 256QAM, each symbol carries 8 bits of information, fully exploiting the transmission rate potential when the channel conditions are good; The data to be transmitted is split into two paths, and horizontal polarization and vertical polarization processing are carried out respectively. After the horizontal polarization signal is processed by a dedicated modulator, it is transmitted by the horizontal polarization antenna; after the vertical polarization signal is processed by the corresponding modulator, it is synchronously transmitted by the vertical polarization antenna; During the transmission process, the transmitter embeds a high-precision synchronization pilot signal to ensure that the receiver can perceive the change of the modulation method in real time and synchronously update the demodulation parameters; For polarization diversity transmission, calibrate the time and frequency synchronization of the two paths of signals, and utilize the differences in the reflection and refraction characteristics of signals in different polarization directions by the ionosphere to reduce the probability of simultaneous fading; Optionally, dynamically monitor the channel status of two signals. If the signal fading in a certain polarization direction exceeds the threshold, adjust the power distribution of the two signals in real time; Utilize the independent fading characteristics of the ionosphere for horizontally and vertically polarized signals. Through dynamic power distribution, reduce the probability of simultaneous deep fading of signals and significantly improve the link stability; For example, compensate by increasing the signal power in the other polarization direction to ensure reliable demodulation at the receiving end. Combine the decision result of reinforcement learning. If the modulation mode switches frequently, optimize the transmission interval and power of the synchronization pilot. While ensuring the demodulation accuracy, reduce the impact of the switching overhead on the transmission efficiency. Through the above process, achieve efficient switching of the modulation mode and stable transmission of polarization diversity signals, and ensure the reliable transmission of flight attitude data in the ionosphere environment; In summary, this step is a dynamic modulation and polarization diversity transmission method that integrates ionospheric channel sensing, reinforcement learning, and Shannon capacity theory; First, collect ionospheric channel parameters in real time through sensors, construct a reinforcement learning state space that includes real-time channel status and historical modulation strategies. Use 5 modulation modes as the action space, design a reward function that takes into account the transmission rate, bit error rate, and modulation stability. Train the intelligent agent using the deep Q network. When making a decision, first output the initial modulation mode by the deep Q network, and then verify the theoretical bit error rate and channel capacity through the Shannon capacity model. If it exceeds the theoretical limit, select the sub-optimal modulation mode by reducing the order to ensure the balance between transmission reliability and efficiency; The transmitter adjusts the parameters according to the optimal modulation mode, splits the data into horizontally and vertically polarized signals for synchronous transmission, embeds high-precision pilot signals to ensure demodulation synchronization, and dynamically calibrates the power distribution of the two signals. Combine the polarization reflection characteristics of the ionosphere to reduce the fading impact, and at the same time cooperate with satellite communication to achieve a wider coverage, forming a transmission process of monitoring - decision - transmission - feedback; It has the following effects: Mine the modulation strategies in historical ionospheric data through the deep Q network, so as to adapt to the fast-changing channel. Compared with traditional regularized modulation switching, the spectral efficiency is improved; Embed the Shannon limit theory into the reinforcement learning decision-making process. When the high-order modulation recommended by the intelligent agent exceeds the actual capacity of the channel, automatically select the sub-optimal solution by reducing the order to reduce the decision risk driven by pure data and improve the transmission reliability in the scenario where the bit error rate exceeds the standard; Dynamically balance the transmission rate, bit error rate, and modulation stability (switching overhead) through weight coefficients, enabling the system to intelligently switch between high-rate and high-reliability modes and adapt to the full scenario of the ionosphere from stable to extreme perturbation; At the same time, in actual flight attitude control and encrypted video transmission systems, ionospheric communication and satellite communication often work together; Step 3: Monitor and analyze the frequency of satellite signals in real time, calculate the total Doppler frequency shift, adjust the frequency of the received signal accordingly based on the total Doppler frequency shift, and deploy the frequency hopping spread spectrum technology. The transmitter quickly hops and transmits data on different frequency channels according to a predetermined pseudo-random sequence; In this step, after completing encryption modulation and polarization diversity transmission, for the Doppler frequency shift problem in satellite communication, high-precision frequency synchronization and anti-jamming transmission are achieved through a real-time frequency shift compensation algorithm and an adaptive frequency hopping strategy; The specific steps are as follows: Calculate the Doppler frequency shift based on multi-source data fusion; Combined with satellite ephemeris data (orbital velocity ), receiver motion parameters (motion speed , azimuth angle θ), and the ionospheric refraction correction model, calculate the real-time Doppler frequency shift , and the calculation formula is: ; where, is the carrier frequency, c is the speed of light, is the angle between the receiver motion direction and the signal incident direction; Introduce the ionospheric group delay model, and calculate the frequency shift deviation caused by refraction according to the electron density , sum the real-time Doppler frequency shift and the frequency shift deviation to obtain the total Doppler frequency shift; Collect the noise power spectrum of the current frequency band through a spectrum analyzer, and mark the frequency bands with power exceeding the power threshold within 3 consecutive time slots as interference bands; Adaptive adjustment of the frequency hopping bandwidth according to the standard deviation of the total Doppler frequency shift; the frequency hopping bandwidth The calculation formula is: , where, is the minimum frequency modulation bandwidth, is the adjustment coefficient, and bz is the standard deviation of the total Doppler frequency shift; Adopt a chaotic frequency hopping sequence, and the initial seed combines the satellite timestamp and the receiver location hash to enhance the randomness of the sequence and resist predictive interference; Optionally, the chaotic frequency hopping sequence can be generated by the Logistic map (parabolic map); Dynamically adjust the frequency hopping period according to the frequency shift change rate. The specific adjustment process is as follows: Collect the total Doppler frequency shift within a continuous time window (such as every M time slots), calculate the deviation value of the total Doppler frequency shift between adjacent time slots, and calculate the average change rate; Set the high-speed change threshold and the low-speed change threshold; if the average change rate is greater than the high-speed change threshold, it is determined as a high-speed change scenario; if the average change rate is less than the low-speed change threshold, it is determined as a low-speed change scenario; if the average change rate is greater than or equal to the low-speed change threshold and less than or equal to the high-speed change threshold, it is determined as a normal scenario; For the high-speed change scenario, shorten the frequency hopping period, which can be specifically shortened to the minimum period value; For the low-speed change scenario, extend the frequency hopping period, which can be specifically extended to the maximum period value; For the normal scenario, continue to use the current frequency modulation period; Among them, the minimum period value and the maximum period value are set by those skilled in the art according to historical data and experience; After each adjustment, synchronize the new frequency hopping period to the receiving end through the pilot signal to ensure frequency switching synchronization at both the transmitting and receiving ends; Optionally, a hysteresis threshold can be introduced. For example, if it is necessary to downgrade from a high-speed change scenario to a normal scenario, the adjustment can be made only when the downgrading conditions are met in 3 consecutive windows; Perform orthogonal frequency division multiplexing modulation on the encrypted data stream, and each subcarrier is frequency-offset according to the frequency hopping sequence to form a frequency hopping OFDM signal , ; among them, is the modulation symbol, is the symbol period; is the frequency offset term, is the frequency hopping frequency of the k-th subcarrier changing with time t, is the rectangular pulse function, u is the total number of subcarriers, and k is the subcarrier index; Insert dual-frequency pilots in each frequency hopping time slot , , corresponding to horizontal and vertical polarization signals respectively. The receiving end estimates the frequency shift and the frequency hopping phase through the cyclic correlation algorithm; among them, is the part corresponding to the horizontal polarization signal in the dual-frequency pilot, that is, the horizontal polarization pilot signal; is the part representing the vertical polarization signal in the dual-frequency pilot, that is, the vertical polarization pilot signal; Providing more comprehensive information for the receiving end by calculating the dual-frequency pilot in the above way, so as to accurately estimate the frequency shift and the frequency hopping phase, and improve the accuracy and reliability of signal processing; The receiving end feeds back the demodulation bit error rate and the frequency offset estimation error to the transmitting end; When it is detected that the quality of the ionospheric communication link deteriorates, automatically increase the satellite communication frequency hopping power and reduce the frequency hopping interval to ensure seamless switching to the satellite communication dominant mode when the ionosphere fails and maintain end-to-end communication continuity; Optionally, when the signal-to-noise ratio is less than 10 dB, the satellite communication hopping power is automatically increased by 3 dB, and the hopping interval is reduced to 2 ms to ensure seamless switching to the satellite communication dominant mode in case of ionospheric failure and maintain end-to-end communication continuity; In summary, this step calculates the real-time Doppler frequency shift based on multi-source data fusion, introduces the ionospheric group delay model to calculate the refraction frequency shift deviation, sums the two to obtain the total Doppler frequency shift, marks the interference band through a spectrum analyzer, dynamically adjusts the hopping bandwidth according to the standard deviation of the total Doppler frequency shift, uses a chaotic hopping sequence to enhance randomness, divides high-speed, low-speed, and normal scenarios according to the frequency shift change rate, dynamically adjusts the hopping period, synchronizes the receiving end through a pilot signal, performs orthogonal frequency division multiplexing modulation on the encrypted data stream to form a hopping OFDM signal, inserts dual-frequency pilots (horizontal and vertical polarizations), and the receiving end feeds back the demodulation bit error rate and frequency offset error. When the ionospheric link quality deteriorates, the satellite communication hopping power is automatically increased and the hopping interval is reduced to ensure end-to-end communication continuity; It has the following effects: calculating the Doppler frequency shift through multi-source data fusion, combining with the ionospheric model, accurately adjusting the frequency, reducing signal distortion, and improving the quality of the received signal; hopping spread spectrum, chaotic sequence, dynamic hopping bandwidth and period adjustment can effectively avoid the interference band, resist predictive interference, and ensure the reliability of signal transmission; when the ionospheric link quality deteriorates, by increasing the power and reducing the hopping interval, seamless switching to the satellite communication dominant mode can be achieved, reducing communication interruptions; Step four: Use the predicted attitude map comparison mechanism and the motion trajectory prediction model to reverse-infer the complete attitude data and generate control commands, divide the priorities of the control commands, so that the key control data with high priority can skip the clustering process and directly enter the encryption channel, and perform closed-loop control on the flight attitude to meet the real-time control requirements and improve the system data transmission and processing efficiency; In this step, the node graph randomly determines the prediction graph, and the steps of comparing the prediction graph with the corresponding data graph to obtain the accuracy rate include: S301: Query the node graph and predict the motion trajectory according to the end elements and middle elements; S302: Determine the prediction graph according to the motion trajectory and end elements; Step S303: Query and compare the corresponding data graph based on the time sequence to obtain the accuracy rate; When the accuracy rate reaches the preset accuracy threshold, pack the node graph and send it to the requesting end; As another preferred method of this embodiment: Based on the attitude data of multiple historical cycles, construct a time series prediction model, with the input being the past attitude feature vector and the output being the future predicted attitude map; The receiving end will demodulate the incomplete attitude data map in real time with the predicted attitude map Perform contour line matching, calculate the similarity through the dynamic time warping algorithm, and the calculation formula of the similarity is: ; where is the contour vertex of the real-time pose data graph, is the contour vertex of the predicted pose graph, is the Euclidean distance, n is the total number of contour vertices of the real-time pose data graph, is the total number of contour vertices of the predicted pose graph, q is the total number of contour vertices participating in the distance calculation, and y is the y-th contour vertex; If the similarity is higher than the similarity threshold, package the node graph and send it to the requesting end; Based on the data's requirement for flight safety, perform dynamic priority division based on the prediction accuracy rate. The process is as follows: Among them, the priorities include: the highest level, the medium level, and the lowest level; If the accuracy rate is greater than the predicted high accuracy rate limit value, the node graph of this predicted graph and its associated historical pose data are promoted to the highest level. These nodes are preferentially selected as the benchmark during clustering to reduce the real-time data processing delay; specifically, the "distance weighted screening" process of the conventional node graph can be skipped, and the highly reliable node graph can be directly included in the clustering benchmark set to reduce the calculation delay; Trigger the fast packaging mechanism: skip some redundant checks (such as only retaining the check codes of key vertices), and shorten the transmission cycle to ensure that the predicted reliable data is preferentially transmitted; 16QAM modulation can be used to replace the default QPSK to improve the spectral efficiency when the signal-to-noise ratio ≥ 15dB; use the fast mode of elliptic curve encryption (ECC) (key length 128 bits), and the encryption time-consuming <1μs to ensure that the encryption is completed within 10μs after the data is generated; When the accuracy rate is less than the predicted low accuracy rate limit value, lower the priority of the corresponding node graph to the lowest level, extend its transmission cycle, and avoid invalid data occupying the bandwidth; increase the redundancy coding ratio and improve the reliability of data transmission through more check bits to make up for the prediction error; specifically, the transmission cycle can be extended to 5 times that of the medium level and only occupy the remaining bandwidth after all high-level and medium-level data transmissions are completed; skip real-time matching during clustering and directly use the historical optimal node graph (the node with an accuracy rate ≥ 90%) as the centroid to reduce the influence of low-reliable data on the clustering benchmark; use a lightweight encryption algorithm (such as ChaCha20), with a key length of 64 bits, and compress the data 5:1 (retain the main contour and ignore the secondary details) before encryption, and the data volume after compression is reduced by 40%; When the accuracy rate is less than or equal to the high accuracy rate limit value and greater than or equal to the low accuracy rate limit value, it is the medium level; specifically, the contour line similarity between the real-time data graph and the nodes in the pool can be calculated during each clustering, and the top 3 nodes with the highest similarity are selected as the centroid; use a fixed transmission cycle; use the combination of AES-256 encryption + CRC-32 check; In summary, this step constructs a prediction model based on the node graph or historical attitude data to generate a future predicted attitude graph. The similarity between the real-time data graph and the predicted graph is calculated through the contour line matching algorithm, and the data priority (highest level, medium level, lowest level) is dynamically divided according to the accuracy rate. Among them, the highest-level key control data skips the clustering process and directly enters the elliptic curve encryption channel to shorten the processing delay; the medium-level data adopts dynamic clustering and conventional encryption strategies; the lowest-level data reduces the priority, extends the transmission period, and enhances the redundant coding. Finally, through the differential processing of priorities, the closed-loop control of the flight attitude data is realized to meet the real-time and reliability requirements; It has the following effects: The highest-level control data does not need to go through clustering processes such as eigenvalue calculation and data graph generation, and is directly encrypted and transmitted, enabling a rapid real-time response for the UAV's quick maneuvers and reducing the risk of control lag caused by processing delays; Taking the accuracy rate as the core basis for priority division, highly reliable data is automatically promoted to the highest level and the transmission parameters are optimized, while low-reliable data is demoted and error correction is enhanced, optimizing the problem that the traditional fixed-priority scheme cannot adapt to the dynamic changes in data reliability; While ensuring the real-time nature of flight control, it optimizes the intelligent allocation of computing, storage, and bandwidth resources, providing a new data processing paradigm for aerospace scenarios with high dynamics and high reliability requirements, and having great advantages especially in fields such as UAV swarm control and satellite rapid attitude adjustment.

[0011] Embodiment 2: Please refer to Figure 2 As shown, the image transmission system for flight attitude control and encryption provided by the embodiment of the present invention is used to execute the image transmission method for flight attitude control and encryption in the above embodiment, and specifically includes: Multi-level encryption processing module: Using a multi-level asymmetric encryption algorithm system, the data is grouped, an encryption key is assigned to each group of data, and then a secondary encryption operation is performed using elliptic curve encryption technology. At the same time, the data redundancy coding technology is used to embed a specific proportion of redundant information in the encrypted data; The execution process is as follows: The data to be transmitted is packaged according to the request-side configuration, and after eigenvalue calculation, binarization processing, and MeanShift clustering, a data graph containing contour lines is generated; then geometric and statistical features are extracted to form a multi-dimensional vector, and the data characteristics are determined by calculating the weighted mean of the coefficient of variation, and the DBSCAN or SOM clustering algorithm is dynamically switched to adaptively determine the optimal number of node graphs and the reference node graph; Based on the node graph features, a grouping key is generated and adapted to the request-side configuration, the key is secondarily encapsulated using elliptic curve encryption technology, and finally, in combination with the node graph complexity and communication requirements, targeted redundant check codes are embedded in the key areas of the contour lines and the eigenvalue time series; Ionospheric channel processing module: After completing encryption processing, it monitors the channel conditions of the ionosphere in real time, dynamically adjusts the modulation mode of data, switches from basic binary phase shift keying to high-order quadrature amplitude modulation, and uses polarization diversity technology to transmit signals with horizontal polarization and vertical polarization simultaneously; The execution process is as follows: It collects ionospheric channel parameters in real time through sensors, constructs a reinforcement learning state space containing real-time channel states and historical modulation strategies, uses 5 modulation modes as the action space, designs a reward function that takes into account transmission rate, bit error rate, and modulation stability, trains an agent using a deep Q-network. When making a decision, first the deep Q-network outputs the initial modulation mode, then the theoretical bit error rate and channel capacity are verified through the Shannon capacity model. If it exceeds the theoretical limit, the modulation mode is reduced to select a sub-optimal modulation mode to ensure the balance between transmission reliability and efficiency; The sending end adjusts the parameters according to the optimal modulation mode, splits the data into horizontal and vertical polarization signals for synchronous transmission, embeds high-precision pilot signals to ensure demodulation synchronization, and dynamically calibrates the power distribution of the two signals; Satellite signal processing module: It monitors and analyzes the frequency of satellite signals in real time, calculates the total Doppler frequency shift, adjusts the frequency of the received signal accordingly based on the total Doppler frequency shift, and deploys frequency hopping spread spectrum technology. The transmitting end quickly hops and sends data on different frequency channels according to a predetermined pseudo-random sequence; The execution process is: Based on multi-source data fusion, it calculates the real-time Doppler frequency shift, introduces an ionospheric group delay model to calculate the refraction frequency shift deviation, and the sum of the two gives the total Doppler frequency shift. It marks the interference band through a spectrum analyzer, dynamically adjusts the frequency hopping bandwidth according to the standard deviation of the total Doppler frequency shift, uses a chaotic frequency hopping sequence to enhance randomness, divides high-speed, low-speed, and normal scenarios according to the frequency shift change rate, dynamically adjusts the frequency hopping period, and synchronizes the receiving end through pilot signals. It performs orthogonal frequency division multiplexing modulation on the encrypted data stream to form a frequency hopping OFDM signal, inserts dual-frequency pilot signals, and the receiving end feeds back the demodulation bit error rate and frequency offset error. When the ionospheric link quality deteriorates, it automatically increases the satellite communication frequency hopping power and reduces the frequency hopping interval; Instruction priority processing module: It uses a predicted attitude map comparison mechanism and a motion trajectory prediction model to deduce complete attitude data and generate control instructions, divides the priorities of the control instructions, so that key control data with high priority can skip the clustering process and directly enter the encryption channel; The execution process is as follows: Based on the node graph or historical pose data, a prediction model is constructed to generate a future predicted pose graph. The similarity between the real-time data graph and the predicted graph is calculated through a contour line matching algorithm. The data priority is dynamically divided according to the accuracy rate. Among them, the highest-level key control data skips the clustering process and directly enters the elliptic curve encryption channel to shorten the processing delay; the medium-level data adopts dynamic clustering and conventional encryption strategies; the lowest-level data reduces the priority, extends the transmission period, and enhances the redundant coding. Finally, through the differential processing of priorities, the closed-loop control of the flight attitude data is realized to meet the requirements of real-time performance and reliability.

[0012] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art of this industry should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification only illustrates the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A video transmission method applied to flight attitude control and encryption, characterized in that: It includes the following steps: Group the data, assign an encryption key to each group of data, then perform a secondary encryption operation using elliptic curve encryption technology, and at the same time use data redundancy coding technology to embed a specific proportion of redundant information in the encrypted data; After the encryption process is completed, monitor the channel conditions of the ionosphere in real time, dynamically adjust the modulation mode of the data, and use polarization diversity technology to transmit signals with horizontal polarization and vertical polarization simultaneously; Monitor the frequency of the satellite signal in real time, calculate the total Doppler frequency shift, adjust the frequency of the received signal according to the total Doppler frequency shift, and deploy frequency hopping spread spectrum technology. The transmitting end quickly jumps and transmits data on different frequency channels according to a predetermined pseudo-random sequence; Use the predicted attitude map comparison mechanism and the motion trajectory prediction model to reverse infer the complete attitude data and generate control instructions, and divide the priorities of the control instructions, so that the key control data with high priority skips the clustering process and directly enters the encryption channel.

2. The image transmission method applied to flight attitude control and encryption according to claim 1, characterized in that: Data grouping and encryption process, the process is as follows: Convert the data to be transmitted into a data graph, and determine the node graph through pixel point clustering; Generate a hash value based on the eigenvalue group and contour line coordinate sequence of the node graph, and generate a grouping key through a key derivation function; Use the elliptic curve public key of the receiving party to perform secondary encryption on the grouping key, and embed a redundancy check code in the key vertices of the contour line and the eigenvalue time sequence of the encrypted data.

3. The image transmission method for flight attitude control and encryption according to claim 2, wherein: The process of determining the node graph includes dynamic clustering strategy switching; Extract geometric features and statistical features from the data graph containing contour lines to form a comprehensive feature vector containing multi-dimensional features; perform normalization processing on the current batch of feature vector sets, calculate the coefficient of variation of each dimension, and generate a global dispersion index through weighted average; If the global dispersion index ≤ the stationary threshold, it is determined as stationary flight data and the DBSCAN algorithm is used; if ≥ the complex threshold, it is determined as complex maneuver data and the self-organizing mapping SOM algorithm is used.

4. The image transmission method applied to flight attitude control and encryption according to claim 1, characterized in that: Dynamically adjust the modulation mode of the data, the process is as follows: Real-time collect ionospheric channel parameters through sensors built into the transmitting end, and make dynamic decisions in the modulation stage based on the cooperation of reinforcement learning drive and Shannon capacity model.

5. The video transmission method for flight attitude control and encryption according to claim 4, characterized in that: The decision-making process in the modulation stage includes: Construct a reinforcement learning framework including a state space, an action space, and a reward function; Input the real-time state into the trained deep Q network, and select the modulation mode with the largest Q value as the optimal modulation mode; When the theoretical bit error rate of the modulation order recommended by reinforcement learning does not exceed the target bit error rate, calculate the theoretical channel capacity under the optimal modulation mode. If the theoretical channel capacity is less than or equal to the Shannon extreme value, use the optimal modulation mode output by the deep Q network; if the theoretical channel capacity is greater than the Shannon extreme value, select a modulation mode one level lower than the current modulation order from the action space. If the theoretical bit error rate of the newly selected modulation mode still does not exceed the target bit error rate and the theoretical channel capacity is less than or equal to the Shannon extreme value, use this modulation mode; if the conditions are still not met after adjustment, repeat the above operations until a modulation mode that meets the Shannon limit and has a qualified bit error rate is found.

6. The image transmission method applied to flight attitude control and encryption according to claim 5, characterized in that: The decision-making process of the modulation mode also includes: When the theoretical bit error rate corresponding to the modulation order recommended by reinforcement learning exceeds the target bit error rate, automatically fallback to the suboptimal order calculated by the Shannon capacity. Traverse the action space, calculate the theoretical bit error rate for each modulation method, filter out the modulation methods that satisfy the theoretical bit error rate ≤ target bit error rate, construct a set of candidate modulation methods. For each modulation method in the selected modulation method set, calculate its theoretical channel capacity, and sort all the theoretical channel capacities in descending order. The modulation method corresponding to the maximum theoretical channel capacity is used as the optimal modulation method. If there is only one modulation method in the candidate modulation method set, directly output this modulation method as the optimal modulation method.

7. The image transmission method for flight attitude control and encryption according to claim 1, characterized in that: Satellite signal frequency shift compensation and frequency hopping transmission, the process is as follows: Obtain and fuse the satellite orbital velocity, receiver movement velocity, azimuth angle, and ionospheric refraction correction model, and calculate the real-time Doppler frequency shift amount; introduce the ionospheric group delay model, calculate the frequency shift deviation caused by refraction according to the electron density, and sum the real-time Doppler frequency shift amount and the frequency shift deviation to obtain the total Doppler frequency shift amount. Adopt a chaotic frequency hopping sequence, and dynamically adjust the frequency hopping period and bandwidth according to the frequency shift change rate. Insert dual-frequency pilots in the frequency hopping OFDM signal for frequency shift and phase estimation at the receiver.

8. The image transmission method applied to flight attitude control and encryption according to claim 7, characterized in that: Adjust the frequency hopping period, the process is as follows: Collect the total Doppler frequency shift amount within a continuous time window, calculate the deviation value of the total Doppler frequency shift amount between adjacent time slots, and calculate the average change rate. If the average change rate is greater than the high-speed change threshold, it is determined as a high-speed change scenario, and the frequency hopping period is shortened to the minimum period value; if the average change rate is less than the low-speed change threshold, it is determined as a low-speed change scenario, and the frequency hopping period is extended to the maximum period value.

9. The image transmission method for flight attitude control and encryption according to claim 1, characterized in that: Divide the control instructions into priorities, the process is as follows: Construct a time series prediction model based on historical attitude data to generate a predicted attitude map. Calculate the similarity between the real-time data map and the predicted map through the dynamic time warping algorithm, and divide the priorities according to the similarity. If it is greater than the predicted high accuracy limit value, the corresponding data is the highest level; when it is less than the predicted low accuracy limit value, the corresponding data is the lowest level. When it is less than or equal to the high accuracy limit value and greater than or equal to the low accuracy limit value, the corresponding data is the middle level.

10. A video transmission system applied to flight attitude control and encryption, which executes the video transmission method applied to flight attitude control and encryption according to any one of claims 1-9, characterized in that: Specifically include: Multi-level encryption processing module: Group the data, allocate an encryption key for each group of data, then perform a secondary encryption operation using the elliptic curve encryption technology, and at the same time use the data redundancy coding technology to embed a specific proportion of redundant information in the encrypted data. Ionospheric channel processing module: After completing the encryption processing, monitor the channel conditions of the ionosphere in real time, dynamically adjust the modulation method of the data, and use the polarization diversity technology to transmit horizontal polarization and vertical polarization signals simultaneously. Satellite signal processing module: Monitor the frequency of the satellite signal in real time, calculate the total Doppler frequency shift amount, adjust the frequency of the received signal according to the total Doppler frequency shift amount, and deploy the frequency hopping spread spectrum technology. The transmitter quickly hops and transmits data on different frequency channels according to a predetermined pseudo-random sequence. Instruction priority processing module: Using the predicted attitude map comparison mechanism and the motion trajectory prediction model to reverse-infer the complete attitude data and generate control instructions, dividing the priorities of the control instructions, so that the key control data with high priority can skip the clustering process and directly enter the encryption channel.

Images