Unmanned aerial vehicle long-distance aerial photography data encryption transmission method

Abstract

The application discloses a UAV long-distance aerial photography data encryption transmission method, and relates to the cross technical field of information security and UAV remote sensing communication. In the process of UAV flight, the original aerial photography image data collected is pretreated and dynamically blocked by the airborne processing unit, and each data block is allocated a unique identification sequence; a session key is generated in real time in combination with wireless channel state information, UAV flight attitude parameters and ground station relative position information; different intensity encryption processing is respectively performed according to the sensitivity of the data block; after the encrypted data block is subjected to priority scheduling, forward error correction coding and adaptive modulation coding, it is sent to the ground receiving end; after the ground receiving end completes integrity check, a selective retransmission request is initiated, and the airborne processing unit only re-executes the encryption and transmission process for the specified data block; after all the data blocks are received, the ground receiving end recombines according to the unique identification sequence, and decrypts and restores by using the session key, and outputs complete aerial photography image data.

Description

Technical Field

[0001] This invention relates to the interdisciplinary field of information security and UAV remote sensing communication, specifically a method for encrypted transmission of long-distance aerial photography data from UAVs. Background Technology

[0002] With its advantages of mobility, rapid response, low operating costs, and wide coverage, UAV aerial photography technology has been widely applied in many fields, including land surveying, disaster emergency patrols, border and coastal defense control, major engineering surveys, agricultural resource monitoring, and security monitoring of key areas. During operations, UAVs continuously acquire high-resolution images of the ground surface, texture features of ground features, and spatial distribution information through onboard imaging equipment. The efficient and secure transmission of aerial photography data is a core prerequisite for subsequent image stitching, disaster assessment, target identification, and decision-making.

[0003] In many core operational scenarios, real-time transmission of aerial photography data is a rigid requirement for mission execution and cannot be achieved by copying data after the drone lands. In disaster emergency response scenarios, the rescue and handling of sudden emergencies such as forest fires, earthquakes, landslides, floods, and work safety accidents are extremely time-sensitive. As the core means of acquiring first-hand data at disaster sites, drones' aerial images directly determine the efficiency and accuracy of rescue plan formulation, personnel and material deployment, and disaster situation assessment. Aerial photography data must be transmitted back to the ground command center in real time. If data is copied after the drone has finished its flight and landed, the best rescue opportunity will be missed, resulting in irreparable loss of life and property. In border and coastal defense control and security scenarios in key sensitive areas, the core objective of routine inspections of border lines, coastlines, major infrastructure, and classified sensitive areas is to identify security incidents such as illegal border crossings, regional intrusions, and abnormal activities in real time. Only by transmitting aerial images in real time can early warnings be triggered and emergency response procedures initiated immediately. If data is read after landing, the handling of security incidents will be severely delayed, failing to meet the real-time requirements of security control. Third, in ultra-long-distance, long-endurance aerial photography missions, such as large-scale land surveying and marine environmental monitoring, a single UAV flight can last for several hours or even more than ten hours, and the amount of aerial photography data collected in a single mission can reach tens or even hundreds of gigabytes. If only data is copied after landing, on the one hand, the validity of the aerial photography data cannot be verified in real time during the flight, making it difficult to detect problems such as missed shots, blurred images, focus failure, and flight path deviation in a timely manner, which can easily lead to rework. On the other hand, if the UAV loses contact or crashes during flight, all aerial photography data will be lost, resulting in complete mission failure. Real-time transmission can simultaneously complete off-site backup of core data, while monitoring the quality of aerial photography operations in real time, ensuring mission closure. Fourth, in dynamic on-site management scenarios, tasks such as real-time monitoring of the progress of large-scale engineering construction, dynamic inspection of illegal construction, and real-time monitoring of the spread of agricultural pests and diseases require the rapid generation of analysis results based on real-time transmitted aerial photography images to simultaneously guide on-site operations and law enforcement actions. Such scenarios have clear requirements for the real-time nature of data transmission and cannot wait until the flight mission is over before data processing and analysis.

[0004] Therefore, ensuring that the data obtained from aerial photography can be transmitted to the ground station in real time, securely, and stably and processed quickly is a core technical problem that urgently needs to be solved in the field of UAV aerial photography.

[0005] In existing technologies, two main technical solutions are used to achieve real-time transmission of UAV aerial photography data: one is the general wireless image transmission solution. This solution typically uses high-compression lossy compression of aerial images to meet real-time requirements. While this reduces bandwidth usage, it leads to significant loss of spatial resolution, geometric accuracy, and spectral information, failing to meet the high-precision application requirements of surveying, target identification, and emergency analysis. The other is the traditional encrypted data transmission solution. While this solution ensures data security, it still has several insurmountable drawbacks: First, most solutions use a fixed key and fixed encryption strength for full data encryption. However, the computing power and power consumption of the UAV's onboard processing unit are limited by the onboard platform. Full, high-strength encryption incurs significant computational overhead, resulting in a substantial increase in data processing latency, failing to meet the real-time transmission requirements of long-distance aerial photography. Second, fixed-key systems are vulnerable to replay attacks and man-in-the-middle attacks. In long-distance, open wireless transmission links, aerial photography data containing sensitive geographical information and key target features is susceptible to interception and cracking. The significant security risks posed by tampering make it unsuitable for security protection requirements in classified scenarios and sensitive areas. Thirdly, the commonly used transmission strategies of whole-packet retransmission and fixed modulation coding are problematic because the wireless channels of long-distance UAV aerial photography are susceptible to terrain obstruction, electromagnetic interference, and changes in UAV flight attitude, resulting in inherent issues such as bandwidth fluctuations, channel fading, and sudden bit errors. Fixed transmission strategies cannot adapt to dynamic channel changes; partial data packet loss necessitates whole-packet retransmission, significantly consuming limited wireless bandwidth, further exacerbating transmission latency, and even causing real-time transmission link interruptions, making it impossible to maintain stable real-time backhaul in complex environments. Fourthly, existing solutions do not perform end-to-end collaborative optimization of data preprocessing, intelligent segmentation, hierarchical encryption, transmission scheduling, and error correction retransmission. They cannot simultaneously address the three core requirements of real-time transmission of high-precision raw aerial photography data, transmission stability of long-distance wireless links, and transmission security of sensitive aerial photography data. This results in a persistent industry pain point where security and efficiency cannot be balanced in real-time data backhaul for long-distance UAV aerial photography, making it difficult to adapt to the engineering application requirements in complex scenarios.

[0006] To address the aforementioned shortcomings of existing technologies, this invention proposes a method for encrypted transmission of long-distance aerial photography data from unmanned aerial vehicles (UAVs). Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention provides a method for encrypted transmission of long-distance aerial photography data from unmanned aerial vehicles (UAVs), which solves the problems mentioned in the background section.

[0008] To achieve the above objectives, the present invention provides a method for encrypted transmission of long-range aerial photography data from unmanned aerial vehicles (UAVs), comprising the following specific steps: Step 1: During the flight of the UAV, the onboard processing unit preprocesses and divides the raw aerial image data into blocks, dynamically divides the data block granularity according to the complexity of the image content, and assigns a unique identifier sequence to each data block; Step 2: Construct a dynamic key generation model based on the current wireless channel state information, and generate a session key bound to channel characteristics in real time by combining the UAV flight attitude parameters and the relative position information of the ground station; Step 3: Apply different levels of encryption to each data block using a hierarchical encryption strategy. High-sensitivity data blocks are encrypted with high-strength encryption, while low-sensitivity data blocks are encrypted with lightweight encryption. Step 4: The encrypted data blocks are scheduled and encapsulated according to a priority queue, and a fault-tolerant transmission frame structure is generated by combining the forward error correction coding mechanism, and a link status feedback identifier is embedded. Step 5: The adaptive modulation and coding module adjusts the modulation order and coding rate according to the real-time channel quality, maps the transmission frame to the physical layer signal, and sends it to the ground receiver. Step 6: After demodulating and decoding the received signal, the ground receiver verifies the data integrity based on the link status feedback identifier and initiates a selective retransmission request for missing or erroneous data blocks. Step 7: The airborne processing unit responds to the selective retransmission request by re-executing the encryption and transmission process only for the specified data block to avoid retransmitting the entire data. Step 8: After the ground receiver has received all data blocks, it reassembles them according to the unique identifier sequence, decrypts and restores them using the session key, and outputs complete aerial image data.

[0009] Preferably, the preprocessing of the original aerial image data in step 1 includes removing sensor noise, correcting geometric distortion, and adjusting radiometric consistency. The dynamic division of the data block granularity is determined based on a comprehensive assessment of the local texture entropy value and edge density of the image, so that high-information areas are divided into smaller data units.

[0010] Preferably, in step 2, the wireless channel state information includes signal-to-noise ratio, multipath delay spread, and carrier phase offset; the flight attitude parameters include pitch angle, roll angle, and yaw angle; the relative position information is calculated jointly by the airborne global navigation satellite system module and the ground station coordinates; and the dynamic key generation model adopts a pseudo-random sequence generator based on chaotic mapping, whose initial seed is generated by fusing the above-mentioned multi-source parameters.

[0011] Preferably, the criteria for determining the highly sensitive area in step 3 are image sub-regions containing geographic coordinate markers, areas with changes in the outline of ground features, or dense artificial structures. The high-strength encryption mode adopts a multi-round iterative block cipher structure, with each round of operation introducing nonlinear substitution and displacement confusion operations. The lightweight encryption mode adopts a stream cipher structure.

[0012] Preferably, the priority queue construction rule in step 4 is as follows: data blocks corresponding to high-sensitivity areas are assigned high priority, and data blocks that fail to be transmitted for the first time are automatically promoted one level in priority during retransmission. The forward error correction coding mechanism adopts a concatenated coding structure, with the outer code being a cyclic redundancy check code and the inner code being a convolutional code. The transmission frame structure includes a frame header synchronization field, a data payload area, and a tail check area. The link status feedback identifier records the task batch and timing number to which the current frame belongs.

[0013] Preferably, in step 5, the adaptive modulation and coding module supports switching between multiple modulation modes, including binary phase shift keying, quadrature phase shift keying, and quadrature amplitude modulation, and the coding rate is dynamically adjusted according to the predicted channel bit error rate.

[0014] Preferably, the data integrity verification at the ground receiving end in step 6 adopts a two-level verification mechanism. The first level identifies the start position of the valid frame based on the frame header synchronization field, and the second level determines whether the data load is damaged based on the cyclic redundancy check result. The selective retransmission request only contains the unique identifier sequence and time sequence number of the missing or erroneous data block, and does not carry the original data content.

[0015] Preferably, in step 7, after receiving a selective retransmission request, the airborne processing unit first verifies the legitimacy of the request source, then extracts the corresponding data block from the local cache, and re-executes the encryption and transmission process of steps 3 to 5. The local cache retains a predetermined number of recently processed data blocks, and data blocks that exceed the retention period are automatically cleared.

[0016] Preferably, in step 8, the data reconstruction process of the ground receiving end strictly follows the logical order of the unique identifier sequence. If a sequence breakpoint is found, the reconstruction is paused and the missing block is waited for to arrive. The decryption and restoration process corresponds to the reverse of the encryption process. The final output aerial image data maintains the original spatial resolution and spectral characteristics.

[0017] This invention provides a method for encrypted transmission of long-distance aerial photography data from unmanned aerial vehicles (UAVs), which has the following beneficial effects: (1) During system operation, the introduction of a dynamic key generation mechanism based on the fusion of channel state and flight parameters significantly improves the unpredictability of keys and link security, effectively resisting replay attacks and man-in-the-middle attacks against fixed key systems; the adoption of a hierarchical encryption strategy realizes the on-demand allocation of computing resources and security strength, which greatly reduces the overall encryption latency while ensuring the high security level of core geographic information, meeting the real-time requirements of long-distance aerial photography missions; the combination of priority scheduling and selective retransmission mechanism avoids the bandwidth waste caused by traditional full retransmission, and can still maintain efficient and stable data throughput in complex electromagnetic environments; through the collaborative design of adaptive modulation coding and forward error correction, the robustness of physical layer transmission to channel fluctuations is enhanced, and data loss caused by signal fading is reduced; the entire methodology achieves the organic unity of security, efficiency and reliability without increasing additional hardware costs, providing an engineerable encrypted transmission solution for UAV long-distance aerial photography operations.

[0018] (2) This invention combines image content complexity analysis with a dynamic data block mechanism, and adaptively divides aerial images into smaller, more granular parts based on texture entropy and edge density. This allows high-information areas to use finer data block structures, while low-information areas use larger data block structures, thereby reducing the number of data blocks while ensuring the integrity of information in key areas. This approach can improve image transmission efficiency without significantly increasing computational burden and reduce redundant data processing overhead in wireless link environments, enabling UAVs to complete data transmission more efficiently in long-distance aerial photography missions.

[0019] (3) This invention constructs a dynamic key generation model based on wireless channel state, UAV flight attitude parameters, and relative position information of ground stations. It integrates multi-source information such as signal-to-noise ratio, multipath delay spread, carrier phase offset, and flight attitude into initial seed parameters for a chaotic system, achieving deep binding between the session key and real-time communication environment characteristics. This mechanism enables the key generated in each communication process to possess high randomness and dynamism, effectively reducing the risk of key prediction or replay attacks, thereby significantly improving the security and anti-attack capability of aerial photography data in long-distance wireless transmission scenarios.

[0020] (4) This invention employs a synergistic design of hierarchical encryption strategy and adaptive transmission mechanism to implement high-strength block encryption for data blocks in highly sensitive areas and lightweight stream encryption for data blocks in low-sensitivity areas. Combined with priority scheduling, forward error correction coding, and adaptive modulation and coding techniques, it ensures the priority and reliable transmission of critical data in complex wireless environments. Simultaneously, through selective retransmission and block-level reassembly mechanisms, only lost or damaged data blocks are retransmitted, avoiding the duplication of the entire data transmission. This reduces communication latency and bandwidth usage while ensuring data integrity and security, thereby improving the overall stability and transmission efficiency of long-distance aerial photography data links. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the overall technical solution architecture of the UAV long-distance aerial photography data encryption transmission method according to the embodiments of this application; Figure 2 This is a schematic diagram of the core principle framework of the UAV long-distance aerial photography data encryption transmission method according to the embodiments of this application, which is driven by the fusion of wireless channel state, flight attitude and relative position to generate dynamic keys and hierarchical encryption. Figure 3 This is a flowchart illustrating the multi-level air-to-ground interactive logic of the UAV long-distance aerial photography data encryption transmission method based on priority queue scheduling, adaptive modulation and coding, and selective retransmission feedback, according to an embodiment of this application. Detailed Implementation

[0022] 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 of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0023] Example 1 This invention provides a method for encrypted transmission of long-range aerial photography data from unmanned aerial vehicles (UAVs). Please refer to [link / reference]. Figure 1In this embodiment, the preprocessing, segmentation, encryption, encapsulation, and retransmission response of the raw aerial image data are all completed by the onboard processing unit during the UAV's flight. The ground receiving end performs demodulation and decoding, integrity verification, retransmission request feedback, data reconstruction, and decryption restoration. In step 1, the onboard processing unit preprocesses and segments the acquired raw aerial image data, dynamically dividing the data block granularity according to the complexity of the image content, and assigning a unique identifier sequence to each data block. When performing step 1, the high-resolution optical sensor or multispectral imaging device carried by the UAV first acquires the raw aerial image. The raw aerial image data usually contains a large number of raw count values, and due to environmental factors and inherent sensor characteristics, these data have a certain degree of inconsistency. Therefore, the preprocessing process in step 1 first performs a sensor noise removal operation. This operation involves constructing a spatial domain filter, for example, applying a 3x3 or 5x5 window to perform statistical analysis on the pixel values, calculating the mean or median of the pixels within the window, and using this value to replace the center pixel to remove isolated noise points caused by thermal noise of electronic devices or photoelectric conversion processes.

[0024] Subsequently, geometric distortion correction is performed to address attitude disturbances during UAV flight and the optical characteristics of the lens itself. This process involves acquiring the camera's interior orientation elements, including focal length, principal point coordinates, and radial and tangential distortion coefficients. By establishing a mapping model between image point coordinates and object space coordinates, distorted pixel coordinates are mapped to the ideal imaging plane. Bilinear interpolation is then used to resample the grayscale values ​​of non-integer coordinate points, ensuring that the image geometry closely matches the actual contours of ground features. Furthermore, radiometric consistency adjustment is a crucial preprocessing step. This is achieved by linearly mapping gain and offset between images from different exposure times or under different lighting conditions, or by using histogram specification techniques to unify the brightness distribution across the entire image or adjacent flight strips.

[0025] After preprocessing, this embodiment of the invention performs dynamic partitioning of data blocks. This process does not employ traditional fixed-size slicing, but rather intelligently adapts to the complexity of the image content. Specifically, the system calculates the texture entropy and edge density of local image regions in real time. The calculation of texture entropy involves statistically analyzing the probability of each gray level in the image, multiplying each probability value by its corresponding base-2 logarithmic product, and finally taking the negative of the product to characterize the degree of disorder or richness of information in the region. Edge density is determined by performing gradient calculations on the image using the Sobel operator or the Cannibal edge detection operator, calculating the number of pixels with a gradient magnitude exceeding a preset threshold per unit area. When the texture entropy value of an image region is higher than a preset first threshold, and the edge density is higher than a preset second threshold, the region is determined to be a high-information-content region.

[0026] For high-information areas, the system divides them into smaller data units, such as 128x128 pixel blocks; while for flat areas with indistinct texture features, such as water surfaces or uniformly vegetated areas, they are divided into larger data units, such as 512x512 blocks. This dynamic partitioning mechanism ensures that subsequent encryption operations can achieve finer-grained control precision for high-information, high-value areas. Each generated data block is assigned a unique identifier sequence, consisting of a mission number, waypoint index, in-image row and column index, and a random number generated based on the current system's millisecond-level timestamp. This ensures that there are no duplicate identifiers globally throughout the entire aerial photography mission lifecycle, providing a unique index for subsequent data reassembly and deduplication.

[0027] Step 2 involves constructing a dynamic key generation model based on the current wireless channel state information. This model combines the UAV's flight attitude parameters with the relative position information of the ground station to generate a session key bound to channel characteristics in real time. In Step 2, the wireless channel state information is a crucial source entropy for dynamic key generation. The physical layer module in the UAV communication link monitors and extracts the signal-to-noise ratio (SNR), multipath delay spread, and carrier phase offset in real time. The SNR is obtained by measuring the ratio of the received pilot signal power to the noise floor power, reflecting the purity of the link. The multipath delay spread is extracted by analyzing the channel impulse response, identifying the time difference between the earliest and latest paths of signal arrival, reflecting the reflection and scattering effects of terrain on radio waves. The carrier phase offset is obtained by tracking the phase change of the received signal through a phase-locked loop, acquiring the phase drift caused by Doppler shift or local oscillator frequency deviation.

[0028] Meanwhile, flight attitude parameters, including the UAV's pitch, roll, and yaw angles in the navigation coordinate system, are acquired by the onboard inertial measurement unit at high frequency and processed using a Kalman filter algorithm to eliminate measurement errors caused by sensor drift. Relative position information is calculated using latitude, longitude, and altitude obtained from the onboard global navigation satellite system module, combined with pre-stored precise geographic coordinates of the ground station. The specific calculation process employs a geocentric coordinate system transformation model to calculate the three-dimensional straight-line distance and relative azimuth between the UAV and the ground station.

[0029] The dynamic key generation model employs a pseudo-random sequence generator based on chaotic mapping. Its core logic lies in leveraging the extreme sensitivity of chaotic systems to initial conditions. The system normalizes the aforementioned multi-source parameters—signal-to-noise ratio, multipath delay spread, carrier phase offset, pitch angle, roll angle, yaw angle, and relative distance—and then generates a high-precision floating-point number using a weighted fusion algorithm. This floating-point number serves as the initial seed parameter for the chaotic mapping system. The chaotic mapping system performs multiple rounds of iterative computation using either a logical mapping model or a Lorentz mapping model. In each iteration, the current system state value is mapped using a nonlinear function to generate the next state value. After thousands of iterations, the unstable outputs from the previous stages are discarded, and the subsequently generated chaotic sequence is extracted. Due to the inherent unpredictability and uniqueness of channel characteristics and flight attitude during long-distance aerial photography, the generated chaotic sequence exhibits extremely high randomness and anti-interception capabilities. Finally, a quantization operator converts the floating-point chaotic sequence into a binary bit string of a specified width, which serves as the session key for the current communication task.

[0030] Step 3 employs a hierarchical encryption strategy, applying encryption operations of varying strengths to each data block. Data blocks corresponding to highly sensitive areas utilize a high-strength encryption mode, while those corresponding to low-sensitivity areas utilize a lightweight encryption mode. When executing step 3, the sensitivity of the data blocks must first be automatically determined. The criteria for determining highly sensitive areas primarily include three dimensions: first, whether the geographical location corresponding to the data block contains preset sensitive coordinate markers; second, whether the image content exhibits significant changes in the outline of features, such as coastlines or border lines; and third, whether it contains densely populated areas of artificial structures, such as building clusters or transportation hubs. This is achieved through target detection algorithms on the segmented image.

[0031] For data blocks identified as highly sensitive areas, a high-strength encryption mode is activated. This mode employs a multi-round iterative block cipher structure, with core operations including nonlinear substitution, row shift obfuscation, column obfuscation, and round key addition. Nonlinear substitution is performed using a carefully designed substitution box (S-box), mapping the input byte to another completely unrelated byte to disrupt the linear statistical properties of the data. Row shift obfuscation involves cyclically shifting each row of the block matrix to the left, with the shift amount increasing with the row number. Column obfuscation achieves deep bit diffusion by treating each column as a polynomial over a finite field and multiplying it by a fixed polynomial. Round key addition involves XORing the current round's subkey with the data block. These multiple rounds of combined operations ensure that any minute changes in the plaintext are propagated throughout the entire ciphertext block.

[0032] For data blocks identified as low-sensitivity areas, a lightweight encryption mode is enabled to conserve UAV computing resources and reduce processing latency. This mode employs a stream cipher structure, which works by generating a long bitstream key stream using the session key. The key stream generation process is implemented through a feedback shift register structure, where the register's state changes and outputs one bit at a time. Compared to the high-strength mode, the lightweight mode omits complex S-box substitution and multi-dimensional matrix transformations, simply merging the key stream with the original data block bit by bit using a simple XOR operation. This approach significantly improves data processing speed while maintaining basic security strength, making it particularly suitable for processing large amounts of background data (such as forests and oceans) during aerial photography.

[0033] In step 4, the encrypted data blocks are scheduled and encapsulated according to a priority queue. A fault-tolerant transmission frame structure is generated using a forward error correction coding mechanism, and a link status feedback identifier is embedded. In step 4, the construction of the priority queue follows strict rules. Data blocks corresponding to highly sensitive areas are assigned the highest priority and enter the head of the transmission buffer queue. Simultaneously, the system monitors the transmission status of each data block in real time. If a data block fails to transmit on its first attempt due to channel fluctuations, its priority will be automatically increased by one level when it enters the retransmission process, ensuring that critical core data can reach the ground end with the shortest possible delay.

[0034] To address sudden interference and signal fading in long-distance transmission, this invention introduces a forward error correction coding mechanism. This mechanism employs a concatenated coding structure, including an outer code and an inner code. The outer code uses a cyclic redundancy check (CRC) code, which works by appending a redundant parity bit after the original data block. This parity bit is obtained by treating the original data as a long polynomial and dividing it by a preset generator polynomial; the remainder is the parity value. The inner code uses a convolutional code, which maps the input bit sequence into a longer output sequence using a set of shift registers with memory and adders. Each bit in the output sequence is related not only to the current input but also to several previous input bits, thus providing extremely strong error correction capabilities.

[0035] The transmission frame structure comprises a frame header synchronization field, a data payload area, and a tail check area. The frame header synchronization field contains a specific bit sequence used by the receiver for clock recovery and frame start-of-position identification. The data payload area carries encrypted and error-corrected data blocks. The tail check area contains integrity verification information for the current frame. Furthermore, a link state feedback identifier is embedded in the frame structure. This identifier records the task batch number to which the current frame belongs and its timing number in the sequence, used by the ground receiver to assess the current packet loss rate and channel quality in real time.

[0036] In step 5, the adaptive modulation and coding module adjusts the modulation order and coding rate based on real-time channel quality, mapping the transmission frame to a physical layer signal and sending it to the ground receiver. In step 5, the adaptive modulation and coding module is crucial for connecting the logical link layer and the physical channel. This module supports dynamic switching between various modulation schemes, including binary phase-shift keying (BPSK), quadrature phase-shift keying (QPSK), and quadrature amplitude modulation (QAM) of different orders. The modulation scheme selection is based on the predicted channel bit error rate (BER). The airborne processing unit receives channel quality indicators from the ground in real-time and performs comprehensive prediction based on collected flight status data. When the predicted BER is below a preset threshold, the module switches to a higher-order modulation scheme and a higher coding rate combination; when the predicted BER increases, the module lowers the modulation order and reduces the coding rate, increasing error correction redundancy. The modulated signal then undergoes digital-to-analog conversion and RF front-end power amplification before being transmitted to the ground receiver via a high-gain antenna.

[0037] In step 6, after demodulating and decoding the received signal, the ground receiver verifies data integrity based on the link state feedback identifier and initiates selective retransmission requests for missing or erroneous data blocks. After capturing the radio frequency signal, the ground receiver first converts it to an intermediate frequency or baseband signal using a low-noise amplifier and a downconverter. Subsequently, the demodulator performs symbol synchronization and carrier recovery, mapping the continuous waveform back to the digital bit stream. The decoder then performs error correction and verification according to the internal and external code protocols corresponding to the transmitter.

[0038] Data integrity verification employs a two-level verification mechanism. The first level of verification is based on the frame header synchronization field, using sliding correlation operations to identify the start boundary of a valid frame. If the synchronization field cannot be detected, or the matching degree of the synchronization field is below a threshold, the frame is considered lost. The second level of verification calculates the checksum corresponding to the received data payload based on the cyclic redundancy check (CRC) result and compares it with the checksum carried in the frame tail. If the two are not equal, it indicates that an irreparable bit error occurred during data transmission. In this case, the ground receiver sends a selective retransmission request via the reverse link. This request contains a compact control packet that carries only the unique identifier sequence of the missing or erroneous data block and its timing number in the task flow, without carrying any original image content, thus maximizing bandwidth savings on the reverse channel.

[0039] In step 7, the airborne processing unit responds to the selective retransmission request by re-executing the encryption and transmission process only for the specified data block, avoiding the complete retransmission of all data. When the airborne processing unit receives a selective retransmission request, it first verifies the legitimacy of the request source by checking whether its sequence number conforms to the current task logic. After successful verification, the processor directly accesses the local cache module. The local cache module adopts a circular buffer design, retaining a predetermined number of recently processed and transmitted data blocks. The system quickly retrieves the target data block using a unique identifier sequence. If the data block is still in the cache, it is retrieved. Subsequently, the system re-executes steps 3 to 5 to generate new encryption parameters based on the wireless channel state corresponding to the retransmission time, and completes the data block transmission using the current adaptive modulation and coding strategy. For data blocks that have exceeded the retention period and have been removed from the cache, the system no longer responds to the retransmission request, or the upper-layer protocol performs exception handling.

[0040] In step 8, after the ground receiver has received all data blocks, it reassembles them based on the unique identifier sequence and decrypts and restores them using the session key, outputting complete aerial image data. In step 8, the ground receiver establishes a reassembly buffer. As data blocks arrive continuously, the reassembly logic places them in the corresponding memory location based on the unique identifier sequence carried by each block.

[0041] The data reassembly process strictly follows the logical order. Due to network latency or retransmission mechanisms, the arrival order of data blocks may differ from the sending order. The reassembly logic continuously monitors the continuity of the sequence. If a breakpoint is detected in the sequence, i.e., a data block with a certain number has not yet arrived, the reassembly process will pause at that point, and the subsequently received data blocks will be temporarily stored in the processing queue until the missing block arrives successfully through the retransmission mechanism. After all the predetermined data blocks have been correctly received and arranged, the system calls the decryption module.

[0042] The decryption and restoration process is the strict reverse of the encryption process. For high-strength encrypted blocks, reverse cyclic shifting, reverse S-box replacement, and corresponding round key XOR are used; for lightweight encrypted blocks, the same session key generation algorithm is used to generate a synchronized key stream for XOR elimination. Since encryption and decryption share a dynamic session key generated based on channel characteristics, and the algorithm logic is completely symmetrical, it ensures that the restored data is completely consistent with the original preprocessed data at the bit level. Finally, the system stitches the reconstructed data blocks into a complete image matrix, outputting aerial images with the original spatial resolution, spectral characteristics, and geographic coordinate information.

[0043] Example 2 Building upon Example 1, this example further refines the encrypted transmission scheme for long-range aerial photography data from unmanned aerial vehicles (UAVs) in complex urban environments. To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0044] In complex urban environments, wireless channels exhibit significant multipath fading and shadowing effects due to the shielding, reflection, and diffraction of electromagnetic waves by buildings. To address this, the extraction accuracy of wireless channel state information is further improved in step 2. The carrier phase offset monitoring frequency is increased to the sub-millisecond level to capture the rapidly changing channel environment caused by UAVs navigating between buildings. When fusing flight attitude parameters, the dynamic key generation model incorporates the second derivative information from the accelerometer to sense instantaneous vibrations. This microscopic vibration characteristic is added as a randomness enhancement factor to the initial seed of the chaotic mapping, resulting in a more time-varying session key.

[0045] In step 3, considering the characteristics of urban aerial photography, a color histogram anomaly detection mechanism has been added to the criteria for determining highly sensitive areas. For example, when a specific artificial color distribution or high-contrast edge appears in an image sub-block, the system automatically prioritizes it to the highest level. Simultaneously, the number of iterations in the high-strength encryption mode dynamically scales based on the drone's remaining battery power. When the battery is sufficient, the number of iterations is increased to provide extreme security protection; when the battery is below the warning threshold, the number of iterations is appropriately reduced while ensuring minimum security strength, ensuring the mission can be completed successfully.

[0046] The priority queue scheduling in step 4 introduces a timeliness factor. For aerial photography missions with real-time monitoring requirements, the system calculates the survival time of each data block from acquisition to the current moment. If the survival time is close to the delay threshold set by the application layer, the data block will receive a one-time priority jump. The forward error correction coding mechanism adds an extra layer of simple fountain code based on XOR operation to the concatenated code to handle long-term burst packet loss phenomena common in urban canyons. This multi-level error correction structure ensures that even if some transmission frames are lost, the receiver still has a high probability of recovering the original information through algebraic operations.

[0047] In step 5, the adaptive modulation and coding module adds support for multi-carrier orthogonal frequency division multiplexing (OFDM). The system divides the transmission bandwidth into multiple orthogonal subcarriers and independently assigns a modulation order to each subcarrier based on the signal-to-noise ratio (SNR) distribution on each subcarrier. In urban environments, certain frequencies may be affected by specific interference sources. Through this refined carrier management, interference-affected frequency bands can be avoided, and data can be concentrated on subcarriers with better transmission quality.

[0048] In steps 6 and 7, the feedback logic for selective retransmission requests is optimized. The ground station predicts the potential signal interruption time based on the predicted UAV flight trajectory. Before the interruption occurs, the ground station proactively requests the UAV to back up and transmit high-priority data blocks from its cache. The local cache module corresponding to the onboard processing unit, when detecting excessive storage pressure, prioritizes clearing data blocks that have already undergone multiple retransmission confirmations and have the lowest sensitivity.

[0049] In step 8, the data reconstruction process at the ground receiving end adopts a multi-threaded parallel processing architecture. Due to the massive amount of urban aerial photography data, the decryption module utilizes a high-performance graphics processor to perform large-scale parallel matrix operations, mapping the reverse S-box replacement and column obfuscation operations onto thousands of computing cores, achieving millisecond-level whole-image decryption capability, and ensuring that the command center can acquire clear, encrypted urban aerial photography images in real time.

[0050] Example 3 Based on the above embodiments, this embodiment describes in detail an implementation method for encrypted transmission of long-distance aerial photography data from unmanned aerial vehicles (UAVs) in extreme natural geographical environments (such as high-altitude mountainous areas and large sea areas). To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0051] In high-altitude mountainous areas or vast sea areas, the communication distance between UAVs and ground stations often reaches the physical limit of the link, resulting in significant signal propagation loss and a pronounced tropospheric scintillation effect. In step 1, the preprocessing stage particularly emphasizes radiometric consistency adjustment, utilizing an atmospheric correction model to eliminate the influence of thin clouds and aerosols on image brightness, ensuring that terrain features remain prominent even in unevenly lit environments. The dynamic partitioning of data blocks focuses more on terrain undulation features; for areas with drastic slope changes, finer data blocks are used to accurately capture terrain details.

[0052] In step 2, the calculation of relative position information incorporates Earth curvature correction. In long-distance transmission, the line-of-sight range is strictly limited by the Earth's curvature. The system dynamically calculates the current wireless line-of-sight using real-time altitude information obtained from an airborne global navigation satellite system. The dynamic key generation model also incorporates this path loss factor, caused by changes in spatial distance, as one of the generation parameters, ensuring that the key is bound not only to the microscopic characteristics of the channel but also to macroscopic spatial geometric features.

[0053] In step 3, considering the limited bandwidth on long-distance links, the hierarchical encryption strategy is given stronger compression coordination. Before performing high-strength encryption on highly sensitive areas, lossless compression based on wavelet transform is performed first to reduce the amount of data to be encrypted without sacrificing pixel accuracy. The lightweight encryption mode is further simplified by using a pre-set pseudo-random sequence pool and performing fast indexing and XOR based on the sequence number of the current frame to meet the confidentiality requirements of non-critical areas with extremely low computational overhead.

[0054] Example 4 This embodiment is an explanation based on Embodiment 1. Please refer to it. Figure 1 Specifically, in step 4, the forward error correction coding mechanism is switched to a low-density parity-check code with long code blocks. This coding method has performance close to the Shannon limit in long-distance, low signal-to-noise ratio environments. A long synchronization pilot band is added to the transmission frame structure to help the ground receiver maintain carrier synchronization under extremely weak signal conditions. The link status feedback identifier not only includes packet loss information but also trend analysis of received signal strength indication, which is used by the UAV to predict the risk of link interruption.

[0055] In step 5, the adaptive modulation and coding module tends to use ultra-low code rate spread spectrum modulation techniques. By spreading the signal energy across a wider spectrum, it achieves signal extraction capabilities far below the noise floor. The adjustment step of the coding rate is refined to cope with slow signal fluctuations caused by atmospheric fading.

[0056] The selective retransmission mechanism in steps 6 and 7 introduces a segmented acknowledgment protocol. Instead of sending feedback after each frame, the ground receiver sends a cumulative acknowledgment for a single data window. This reduces the number of times the reverse link needs to be activated, lowering the UAV's power consumption. The onboard local cache module can recover unacknowledged data blocks from the cache and retransmit them even after a brief power outage and restart during flight.

[0057] In step 8, the ground receiver intelligently fills in and repairs individual unrepairable pixels that appear during the reception process by combining the spatiotemporal correlation of historical aerial photography data. After the decryption and restoration process is completed, the output aerial photography image will be orthorectified using a digital elevation model through a post-processing module, ultimately generating a geographic information product with high geometric accuracy and high security.

[0058] Example 5 For scenarios involving large-scale collaborative aerial photography by drone swarms, this embodiment provides a data encryption transmission scheme based on distributed node forwarding. To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0059] In a cluster collaboration scenario, in step 1, each UAV node not only processes the images it collects but may also act as a relay for other remote nodes. The unique identifier sequence of each data block embeds the source node ID and relay hop count information. In this way, the ground receiver can clearly trace the source and transmission path of each data block.

[0060] In step 2, the dynamic key generation process introduces a collaborative awareness dimension. Neighboring UAVs exchange partial channel statistical characteristics via short-range wireless links. When generating session keys, not only are individual channel parameters considered, but also the overall link consistency within the cluster is taken into account. This multi-node fusion key generation method significantly increases the difficulty of cracking, as attackers must simultaneously acquire multiple dynamically changing channel parameters.

[0061] The hierarchical encryption in step 3 implements a cross-node encryption strategy within the cluster. For highly sensitive data, the encryption process may be completed collaboratively by two drone nodes, with each node holding only a portion of the key. Decryption can only be performed using the complete synthesized key when the data converges to the ground station via a specific path. This distributed encryption mechanism ensures that even if a single drone node is captured, the data block it carries cannot be cracked independently.

[0062] In step 4, the priority queue scheduling incorporates a congestion control mechanism. When a relay node experiences buffer backlog due to excessive forwarding tasks, it notifies the source node to reduce its sending rate via link state feedback, or dynamically adjusts the data flow to select a less busy relay path. The forward error correction coding mechanism employs network coding technology between nodes, algebraically combining data packets from multiple different sources for transmission, further improving the overall cluster transmission efficiency.

[0063] In step 5, the adaptive modulation and coding module implements multi-drone frequency diversity. Adjacent drones use different carrier frequencies or orthogonal code sequences to avoid electromagnetic interference between them. Simultaneously, based on the drone's geometric position within the cluster, its transmit power is dynamically adjusted to achieve energy balance as much as possible while ensuring link connectivity.

[0064] In steps 6 and 7, selective retransmission can be initiated not only by the originating node but also by relay nodes along the path. If a relay node has the target data block in its buffer, it can directly retransmit it in place of the source node, thereby significantly reducing retransmission latency.

[0065] In step 8, the ground receiver needs to process data blocks arriving from multiple channels via different relay paths simultaneously. The reassembly algorithm employs multi-dimensional buffer management, classifying data based on the source node ID. After decryption and restoration, the system seamlessly stitches together aerial images from different UAVs to construct a large-scale, high spatiotemporal resolution real-time map.

[0066] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for encrypted transmission of long-distance aerial photography data from unmanned aerial vehicles (UAVs), characterized in that: Includes the following steps: Step 1: During the flight of the UAV, the onboard processing unit preprocesses and divides the raw aerial image data into blocks, dynamically divides the data block granularity according to the complexity of the image content, and assigns a unique identifier sequence to each data block; Step 2: Construct a dynamic key generation model based on the current wireless channel state information, and generate a session key bound to channel characteristics in real time by combining the UAV flight attitude parameters and the relative position information of the ground station; Step 3: Apply different levels of encryption to each data block using a hierarchical encryption strategy. High-sensitivity data blocks are encrypted with high-strength encryption, while low-sensitivity data blocks are encrypted with lightweight encryption. Step 4: The encrypted data blocks are scheduled and encapsulated according to a priority queue, and a fault-tolerant transmission frame structure is generated by combining the forward error correction coding mechanism, and a link status feedback identifier is embedded. Step 5: The adaptive modulation and coding module adjusts the modulation order and coding rate according to the real-time channel quality, maps the transmission frame to the physical layer signal, and sends it to the ground receiver. Step 6: After demodulating and decoding the received signal, the ground receiver checks the data integrity based on the frame header synchronization field and tail check information in the transmitted frame, and initiates a selective retransmission request for missing or erroneous data blocks. Step 7: The airborne processing unit responds to the selective retransmission request and re-executes the encryption and transmission process only for the specified data block; Step 8: After the ground receiver has received all data blocks, it reassembles them according to the unique identifier sequence, decrypts and restores them using the session key, and outputs complete aerial image data.

2. The method for encrypted transmission of long-distance aerial photography data from unmanned aerial vehicles according to claim 1, characterized in that: The preprocessing process for the raw aerial image data includes: Construct a spatial domain filter, apply a window of preset pixel size to perform statistical analysis on pixel values, calculate the mean or median of pixels within the window and use this value to replace the center pixel in order to eliminate sensor noise; The camera's interior orientation elements are obtained, including focal length, principal point coordinates, radial distortion coefficient, and tangential distortion coefficient. By establishing a mapping relationship model between image point coordinates and object space coordinates, the distorted pixel coordinates are mapped to the ideal imaging plane, and the gray values ​​of non-integer coordinate points are resampled using a bilinear interpolation algorithm. Perform radiometric consistency adjustment by linearly mapping the gain and offset of images at different exposure times or under different lighting conditions, or use histogram specification techniques to make the brightness distribution of the entire image or adjacent flight strip images tend to be uniform.

3. The method for encrypted transmission of long-distance aerial photography data from unmanned aerial vehicles according to claim 1, characterized in that: The process of dynamically dividing the data block granularity based on the complexity of image content includes: Real-time calculation of texture entropy and edge density in local image regions; The process of calculating the texture entropy involves statistically analyzing the probability of each gray level appearing in the image, accumulating the product of each probability value and its corresponding base-2 logarithmic operation result, and finally taking the opposite number. Edge density is determined by performing gradient calculations on the image to count the number of pixels with a gradient magnitude exceeding a preset threshold per unit area. When the texture entropy value of a certain image region is higher than a preset first threshold and the edge density is higher than a preset second threshold, the region is determined to be a high information region and is divided into a preset first size data block; For flat regions with texture features below the first threshold, they are divided into preset second-size data blocks, where the pixel area of ​​the second-size data block is larger than the pixel area of ​​the first-size data block.

4. The method for encrypted transmission of long-distance aerial photography data from unmanned aerial vehicles according to claim 1, characterized in that: The process of generating session keys in real time includes: Extract wireless channel state information, which includes signal-to-noise ratio, multipath delay spread, and carrier phase offset. The flight attitude parameters of the UAV are obtained, including pitch angle, roll angle and yaw angle in the navigation coordinate system, and processed by Kalman filter algorithm; The relative position information is obtained by using the latitude, longitude and altitude obtained by the airborne global navigation satellite system module and the pre-stored geographic coordinates of the ground station through a geocentric coordinate system transformation model to calculate the three-dimensional straight-line distance and relative azimuth between the UAV and the ground station; After normalizing the signal-to-noise ratio, multipath delay spread, carrier phase offset, pitch angle, roll angle, yaw angle, and three-dimensional straight-line distance, a floating-point number is generated using a weighted fusion algorithm as the initial seed parameter for the chaotic mapping system. The chaotic mapping system is used to perform multiple rounds of iterative calculations to extract the chaotic sequence after iteration. The floating-point chaotic sequence is then converted into a binary bit string of a specified width using a quantization operator, which serves as the session key.

5. The method for encrypted transmission of long-distance aerial photography data from unmanned aerial vehicles according to claim 1, characterized in that: The criteria for determining highly sensitive areas include a combination of the following dimensions: Determine whether the geographic location corresponding to the data block contains preset sensitive coordinate markers; Determine whether the image content shows significant changes in the outline of features, including coastlines or borders; The image is determined to contain areas of dense artificial structures, which are identified by a target detection algorithm on the segmented image.

6. The method for encrypted transmission of long-distance aerial photography data from unmanned aerial vehicles according to claim 1, characterized in that: The high-strength encryption mode employs a multi-round iterative block cipher structure, and its core operations include: Non-linear substitution maps one byte of input to another using a substitution box; Row shift obfuscation involves cyclically shifting each row of the grouped matrix to the left. Column obfuscation treats each column as a polynomial over a finite field and multiplies it by a fixed polynomial to achieve deep diffusion of bits. Round key addition involves performing an XOR operation between the current round's subkey and the data block; The lightweight encryption mode employs a stream cipher structure, using the session key to generate a key stream. The key stream generation process is implemented through a feedback shift register structure. At each moment, the state of the register changes and outputs a bit. The key stream is then merged bit by bit with the original data block through an XOR operation.

7. The method for encrypted transmission of long-distance aerial photography data from unmanned aerial vehicles according to claim 1, characterized in that: The construction rules for the priority queue include: Assign high priority to the data blocks corresponding to the highly sensitive areas and place them at the head of the send buffer queue; The transmission status of each data block is monitored in real time. If a data block fails to be transmitted for the first time due to channel fluctuations, its priority is automatically increased by one level when it enters the retransmission process. The transmission frame construction process includes forward error correction coding and integrity verification. The forward error correction coding uses convolutional codes, which map the input bit sequence to the output sequence through shift registers with memory function and adders. The integrity verification uses cyclic redundancy check codes, which obtain the check value by treating the original data as a polynomial and dividing it by a preset generator polynomial.

8. The method for encrypted transmission of long-distance aerial photography data from unmanned aerial vehicles according to claim 1, characterized in that: The operation logic of the adaptive modulation and coding module includes: The system receives channel quality metrics from the ground station in real time and combines them with perceived UAV attitude data to predict the channel bit error rate. When the predicted bit error rate is lower than the preset bit error threshold, the system switches to a combination of high-order modulation and high coding rate. The high-order modulation includes 64th-order quadrature amplitude modulation or 256th-order quadrature amplitude modulation. When the predicted bit error rate is higher than the bit error threshold, the modulation order is reduced to quadrature phase shift keying or binary phase shift keying, and the coding rate is reduced to increase the error correction redundancy. After the transmission frame is mapped to the physical layer signal, it undergoes digital-to-analog conversion and RF front-end power amplification before being transmitted through a high-gain antenna.

9. The method for encrypted transmission of long-distance aerial photography data from unmanned aerial vehicles according to claim 1, characterized in that: The data integrity verification adopts a two-level verification mechanism: The first level of verification is based on the frame header synchronization field in the transmission frame structure. The starting boundary of a valid frame is identified through sliding correlation operation. If the synchronization field cannot be detected or the matching degree is lower than the threshold, the frame is determined to be lost. The second-level verification is based on the cyclic redundancy check result. It calculates the check value corresponding to the received data payload and compares it with the check value carried at the end of the frame. If they are not equal, the data is determined to be corrupted. The selective retransmission request is a control packet that carries a unique identifier sequence of missing or erroneous data blocks and a timing number in the task flow, but does not carry the original image content. Upon receiving a selective retransmission request, the airborne processing unit accesses the local cache module, which employs a circular buffer design, retrieves the target data block using a unique identifier sequence, and re-executes the encryption and transmission process.

10. The method for encrypted transmission of long-distance aerial photography data from unmanned aerial vehicles according to claim 1, characterized in that: The recombination and restoration process includes: The ground receiver establishes a reassembly buffer, and the reassembly logic places each data block in the corresponding memory location based on the unique identifier sequence carried by each data block. The continuity of the sequence is continuously monitored. If a breakpoint is found in the sequence, the reassembly process at the current breakpoint is paused, and the received subsequent data blocks are temporarily stored in the waiting queue until the missing data block arrives successfully through the retransmission mechanism. After all the pre-defined data blocks have been correctly received and arranged, the decryption module is called to perform the reverse operation of the encryption process. For high-strength encrypted blocks, reverse cyclic shift, reverse replacement box replacement, reverse column obfuscation, and corresponding round key XOR are used; for lightweight encrypted blocks, the same session key generation algorithm is used to generate a synchronized key stream for XOR elimination.

Images