Satellite portable station audio and video data compression method and system

Abstract

The invention belongs to the technical field of satellite communication, and relates to a satellite portable station audio and video data compression method and system, and the method comprises the steps: generating a multi-dimensional environment label containing communication stability and disaster grade features; evaluating bandwidth quality based on an environment label, performing adaptive noise reduction and dynamic key encryption on audio and video data, and extracting time-space key features to generate a compressed and encrypted code stream; dynamically dividing the compressed code stream into a core layer and an enhancement layer to form a priority codebook; according to the multi-network congestion state and the satellite switching cost, a codebook level transmission path is allocated, and a destroy-resistant priority label is marked for core data; the redundancy fragments are packaged by using self-adaptive erasure codes, and high-priority data are sent to a low-delay link through a hardware encryption unit; and the block chain reconstructs the synchronous audio and video stream after verifying the fragment integrity. According to the method, the problem that instantaneous interruption caused by shielding or interference cannot be handled by a single satellite link architecture is solved, and the multi-path switching response delay is high.

Description

Technical Field

[0001] The invention belongs to the technical field of satellite communications and relates to a satellite portable station audio and video data compression method and system. Background Art

[0002] In emergency scenarios such as natural disasters and accident rescue, real-time transmission of on-site audio and video data is crucial for command and dispatch. However, complex terrain conditions, volatile satellite channel conditions, and the impact of high concurrent data traffic often cause significant fluctuations in transmission link bandwidth, making traditional fixed compression algorithms difficult to adapt to dynamic network environments. Furthermore, electromagnetic interference and cyberattacks at disaster sites make conventional encryption methods vulnerable to cracking, posing a risk of critical data leakage and significantly impacting the timeliness and security of rescue decisions.

[0003] Traditional approaches often employ strategies that combine preset compression ratios, static redundant coding, and a single satellite communication link. For example, these strategies utilize H.264 standard encoding with Rayleigh fading compensation algorithms to enhance error resilience, hard-coded data priorities, or lightweight encryption using pre-shared keys. These solutions sacrifice compression efficiency or introduce fixed redundant packets to ensure basic transmission requirements, meeting conventional communication requirements in stable channel environments. Some systems attempt to enhance robustness through multipath diversity transmission, but these systems lack dynamic optimization mechanisms for sudden disaster scenarios.

[0004] Based on the above problems, traditional solutions have significant defects in highly dynamic emergency scenarios. The architecture that relies solely on satellite links cannot cope with instantaneous interruptions caused by obstruction or interference, and the multipath switching response delay is high. Summary of the Invention

[0005] In a first aspect, the present invention provides a method for compressing audio and video data of a satellite portable station, which adopts the following technical solution:

[0006] A method for compressing audio and video data of a satellite portable station comprises the following steps:

[0007] S1. Obtain the original audio and video data collected by the emergency site sensors, synchronously analyze the current satellite communication environment parameters and disaster types, and generate multi-dimensional environmental feature labels;

[0008] S2: Based on the bandwidth score of the environmental feature label, adaptive noise reduction is performed on the original audio and video data, and a dynamic encryption key sequence is generated. The key spatiotemporal features are extracted and embedded into the chaotic encryption factor to generate a compressed and encrypted joint processing stream.

[0009] S3. Dynamically divide the compressed primary code stream into layers based on the real-time satellite link packet loss rate and Beidou positioning information: the core layer contains key frame outlines and spectrum baseline parameters, and the enhancement layer contains texture details and motion vector data, forming a priority dynamic codebook;

[0010] S4. Utilize a multi-network convergence assessment model, based on the real-time congestion index of the production data network and broadband ad hoc network, and the switching cost of high-throughput satellite transponders, to allocate transmission paths for each layer of the dynamic codebook and assign anti-destruction priority tags to critical data.

[0011] S5. Redundancy encapsulation of the fragmented data set is performed through adaptive forward erasure coding. Combined with the encryption and compression hardware unit built into the satellite portable station, high-priority data streams are directed to the low-latency ad hoc network link, while non-critical data streams are multiplexed on the high-throughput satellite channel.

[0012] S6. The receiving end uses the timestamp to trace back the key chain of the dynamic codebook, and verifies the integrity of each fragment data through blockchain evidence to reconstruct the original semantic fusion audio and video stream.

[0013] A further solution of the present invention generates a multi-dimensional environmental feature label, comprising the following steps:

[0014] The data acquisition terminal composed of high-sensitivity cameras, microphone arrays, and temperature sensors deployed at the emergency site collects the original audio and video data of the scene in real time;

[0015] Obtain the signal transmission delay, transmission bandwidth, carrier-to-noise ratio, and channel bit error rate parameters of the current satellite communication channel, and construct a characteristic sequence of real-time communication parameters through fast Fourier transform;

[0016] Conduct joint analysis of infrared thermal imaging data and terrain vibration spectrum data collected at the disaster site, and use a pre-trained disaster classification model to identify the disaster type;

[0017] The time-frequency characteristics of the original audio and video data, the characteristic sequence of real-time communication parameters and the disaster type are multi-dimensionally weighted fused to generate a multi-dimensional environmental feature label including communication stability factor, disaster level and data redundancy index.

[0018] A further solution of the present invention is to perform multi-dimensional weighted fusion of the time-frequency characteristics of the original audio and video data, the characteristic sequence of the real-time communication parameters, and the disaster type, including the following steps:

[0019] The raw audio and video data is collected by a sensor module consisting of a wide-angle high-definition camera and a circular microphone array at the disaster site, continuously collecting uncompressed sound wave signals and optical image sequences;

[0020] The real-time communication environment parameters are a set of communication channel physical layer parameters measured by an on-site satellite modem, including a measured value of instantaneous available bandwidth, a signal round-trip delay, and a linear quantized value of a carrier-to-noise power ratio;

[0021] The disaster type, which is the recognition result generated by the trained convolutional neural network based on the input on-site infrared spectrum image and the frequency domain characteristics of the foundation micro-vibration, includes preset disaster types such as earthquakes, debris flows, and industrial fires;

[0022] Based on the normalized feature vectors established by three types of indicators, namely audio and video information capacity, communication channel availability and disaster hazard degree, the three types of indicators are mapped into a transmittable label coding group through a weighting factor matrix.

[0023] A further solution of the present invention generates a dynamic encryption key sequence, comprising the following steps:

[0024] Based on the last three digits of the millisecond level of satellite communication delay as the initial value parameter, a 32-bit random number grouping sequence is generated through the iteration of the Logistic chaos equation;

[0025] The iteratively generated chaotic sequence is bit-wise concatenated with the satellite communication delay parameter to generate a dynamic encryption key sequence;

[0026] The chaotic sequence is embedded in the quantization coefficient residual domain of the H.265 coding tree, and a rate-adaptive arithmetic encoder is used to generate jointly compressed and encrypted bitstream data packets.

[0027] A further solution of the present invention, dynamic hierarchical partitioning, includes the following steps:

[0028] The core layer data is filtered using the contour gradient amplitude of the video object and the baseline slope of the audio spectrum;

[0029] Extracting the enhancement layer texture component according to the spatial distribution density of the motion vector amplitude;

[0030] The layered code rate is dynamically allocated through the regional importance evaluation function, and a layered priority index table is established which is corrected in real time by the satellite position offset, forming a dynamic codebook containing basic reconstruction parameters and incremental supplementary data.

[0031] A further solution of the present invention allocates a transmission path, comprising the following steps:

[0032] Obtain statistics on the TCP retransmission rate of the production data network, the MAC layer conflict probability of the broadband ad hoc network, and the frequency of high-throughput satellite beam switching;

[0033] A network transmission capacity evaluation matrix is ​​constructed based on the real-time congestion index to calculate the throughput loss cost of satellite transponders when switching between adjacent beams.

[0034] The topology lifetime factors of the core layer and enhancement layer data of the dynamic codebook are calculated respectively. The Hungarian algorithm is used to match the optimal transmission path combination, and an anti-destruction priority tag including the satellite grid code, the upper limit of the path switching number and the survival time stamp is added.

[0035] A further solution of the present invention performs redundant encapsulation, comprising the following steps:

[0036] Select encoding parameters of adaptive forward erasure code based on real-time network topology status;

[0037] The core layer uses Galois Field generation matrix operations to generate redundant slices for high-priority data;

[0038] For urgent data with a lifetime less than a preset threshold in the indestructibility priority tag, a multi-point low-latency transmission link is established through the multi-hop routing table of the broadband ad hoc network.

[0039] A further solution of the present invention, blockchain evidence verification, includes the following steps:

[0040] The iteration threshold of the chaotic encryption factor is inferred based on the satellite timing timestamp, and the initial key parameters are restored through the dynamic key tree lookup table.

[0041] The blockchain node queries the pre-stored smart contract under the satellite grid coordinates, and calls the SHA3-256 algorithm to perform on-chain evidence traversal and comparison on the hash summary of the received shard;

[0042] Filter the subset of shards that pass the consistency check to trigger the reconstruction of redundant shards.

[0043] A further solution of the present invention is to restore the initial key parameters through a dynamic key tree lookup table, including:

[0044] Each node represents a key or key fragment. Based on the known key identifier containing the timestamp, iteration number and key version number, the node storing the target key parameters is located in the tree structure.

[0045] In a second aspect, the present invention provides a satellite portable station audio and video data compression system, which adopts the following technical solution:

[0046] A satellite portable station audio and video data compression system includes the following modules:

[0047] The data acquisition module is used to collect raw audio and video data, synchronously analyze the current satellite communication environment parameters and disaster types, and generate multi-dimensional environmental feature labels;

[0048] The pre-processing module is used to score the bandwidth of environmental feature tags, adaptively reduce noise on raw audio and video data, generate dynamic encryption key sequences, and generate compressed and encrypted jointly processed bitstreams;

[0049] The dynamic codebook partitioning module is used to extract the data of the core layer and the enhancement layer in layers according to the packet loss rate and satellite positioning information to form a priority dynamic codebook;

[0050] Path allocation module, used to allocate transmission paths and add anti-destruction priority tags through the multi-network fusion evaluation model;

[0051] Redundant encapsulation module, which uses adaptive forward erasure coding to generate redundant fragments, directing high-priority data streams to low-latency ad hoc network links, and multiplexing non-critical data streams into high-throughput satellite channels;

[0052] The verification and reconstruction module is used to trace back the key chain and verify the integrity of the fragments, and generate a fused stream through audio and video synchronization.

[0053] In summary, the present invention has the following beneficial technical effects:

[0054] 1. Significantly improve the reliability and real-time performance of emergency communication systems through multi-dimensional dynamic perception and intelligent adaptation. By integrating satellite communication parameters and disaster characteristics with audio and video data from the disaster site, a compression and encryption strategy adapted to current channel conditions can be dynamically generated, effectively addressing technical challenges such as large bandwidth fluctuations and multiple interference sources in complex environments.

[0055] 2. A hybrid transmission mechanism based on chaotic encryption and dynamic layering greatly enhances data anti-interception and anti-destruction capabilities. Satellite delay parameters are used to drive chaotic sequences to generate dynamic keys, which are then embedded in the coding domain. Combined with the priority transmission of key frame profile data in the core layer, this ensures that critical information can be transmitted securely and reliably even under harsh channel conditions.

[0056] 3. The deep integration of blockchain evidence storage and forward erasure codes has created a new model for trusted transmission of emergency data. With the help of a dynamic key chain bound to a timestamp and real-time verification of a distributed ledger, the traceability and tamper-proofness of data transmission are ensured. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. The drawings are used to provide a further understanding of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0058] Figure 1 Disclosed is a flow chart of a method for compressing audio and video data of a satellite portable station.

[0059] Figure 2 Disclosed is a structural diagram of an audio and video data compression system for a satellite portable station. DETAILED DESCRIPTION

[0060] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0061] The following is combined with Figure 1-Figure 2 The preferred embodiments of the present invention are described in detail.

[0062] Refer to the attached Figure 1 The present invention proposes a satellite portable station audio and video data compression method, comprising the following steps:

[0063] S1. Obtain the original audio and video data collected by the emergency site sensors, synchronously analyze the current satellite communication environment parameters and disaster types, and generate multi-dimensional environmental feature labels;

[0064] S2: Based on the bandwidth score of the environmental feature label, adaptive noise reduction is performed on the original audio and video data, and a dynamic encryption key sequence is generated. The key spatiotemporal features are extracted and embedded into the chaotic encryption factor to generate a compressed and encrypted joint processing stream.

[0065] S3. Dynamically divide the compressed primary code stream into layers based on the real-time satellite link packet loss rate and Beidou positioning information: the core layer contains key frame outlines and spectrum baseline parameters, and the enhancement layer contains texture details and motion vector data, forming a priority dynamic codebook;

[0066] S4. Utilize a multi-network convergence assessment model, based on the real-time congestion index of the production data network and broadband ad hoc network, and the switching cost of high-throughput satellite transponders, to allocate transmission paths for each layer of the dynamic codebook and assign anti-destruction priority tags to critical data.

[0067] S5. Redundancy encapsulation of the fragmented data set is performed through adaptive forward erasure coding. Combined with the encryption and compression hardware unit built into the satellite portable station, high-priority data streams are directed to the low-latency ad hoc network link, while non-critical data streams are multiplexed on the high-throughput satellite channel.

[0068] S6. The receiving end uses the timestamp to trace back the key chain of the dynamic codebook, and verifies the integrity of each fragment data through blockchain evidence to reconstruct the original semantic fusion audio and video stream.

[0069] In one embodiment of the present invention, step S1 includes the following steps:

[0070] Specifically, a data acquisition terminal consisting of a high-sensitivity camera, microphone array, and temperature sensor deployed at the emergency site collects raw audio and video data in real time. The time domain waveform and spectral characteristics of the collected signal are pre-verified. The signal transmission delay, transmission bandwidth, carrier-to-noise ratio, and channel bit error rate parameters of the current satellite communication channel are also obtained. A characteristic sequence of real-time communication parameters is constructed using a fast Fourier transform.

[0071] A joint analysis of infrared thermal imaging data and terrain vibration spectrum data collected at the disaster site was performed, using a pre-trained disaster classification model to identify the disaster type. The time-frequency characteristics of the original audio and video data and the characteristic sequences of satellite communication parameters were spatially aligned with the disaster type. Through multi-dimensional weighted fusion at the feature coding layer, a multi-dimensional environmental feature label was generated, including a communication stability factor, disaster level, and data redundancy index.

[0072] The raw audio and video data is collected at the disaster site by a sensor module consisting of a wide-angle, high-definition camera and a circular microphone array, continuously collecting uncompressed acoustic signals and optical image sequences. Satellite communication environment parameters are a set of communication channel physical layer parameters measured by on-site satellite modems. These parameters include the measured instantaneous available bandwidth, round-trip signal delay, and the linearly quantified carrier-to-noise power ratio.

[0073] Disaster type: A trained convolutional neural network classifies the input on-site infrared spectral imagery and ground micro-vibration frequency domain characteristics to generate identification results. This includes preset disaster types such as earthquakes, debris flows, and industrial fires. Multidimensional environmental feature labels are normalized feature vectors based on three indicators: audio and video information capacity, communication channel availability, and disaster hazard severity. These three indicators are mapped into a transmittable label code group using a weighting factor matrix.

[0074] For example, assume that the sensor array deployed at the petrochemical fire scene collects explosion sound wave data and infrared video streams, and simultaneously monitors the satellite communication channel to obtain an instantaneous bandwidth of 12Mbps, a delay of 380ms, and a bit error rate of 1.2×10 -3 The flame morphological features are convolved using the preset disaster classification model. The flame image is normalized to 224×224 pixels, histogram equalization is performed to enhance the contrast, and a 3×3 convolution kernel is used to extract the three-level features:

[0075] Primary features include edges and textures (64 feature maps); intermediate features include flame morphological features (128 feature maps); advanced features include combustion dynamics features (256 feature maps); the final fully connected layer outputs a probability vector, which is activated by the Softmax function to determine that the disaster type code is DT3 industrial fire.

[0076] The audio signal is transformed through a Mel-spectrogram to extract a 128-dimensional feature vector. Communication parameters are standardized, and a feature encoder is used to perform a multi-dimensional weighted fusion of speech features, communication parameters, and fire severity. This generates a multi-dimensional environmental feature label in the FFT-ENC213:M12-DT3 format. FFT represents the frequency domain feature encoding, ENC213 is the communication parameter index, M12 represents the multimedia data size, and DT3 is the disaster severity identifier.

[0077] In one embodiment of the present invention, the pre-trained disaster type classification model includes the following steps:

[0078] During the pre-training phase, a sample database of typical disasters, including earthquakes, debris flows, and industrial fires, was constructed. Disaster characteristic data was collected using infrared thermal imagers and three-dimensional geological vibration sensors deployed in the disaster recovery simulation laboratory and at historical disaster sites. The infrared thermal imagers captured temperature field distribution data at a rate of 20 frames per second, generating infrared spectral image sequences with a resolution of 640×480. The geological vibration sensors recorded surface microvibration signals at a sampling rate of 10kHz. A Fourier transform was then performed to generate a spectral energy distribution histogram for the 0-5kHz frequency band.

[0079] Emergency experts label each sample data set with a disaster type label based on the morphology of the radiation temperature anomaly area and the fundamental vibration frequency characteristics. These labels serve as a training set to construct and train a disaster classification model. The collected infrared thermal imaging data and terrain vibration spectrum data serve as input, and pre-set disaster types such as earthquakes, debris flows, and industrial fires serve as output. Disaster types are annotated by emergency experts, including earthquakes labeled as DT1 (DT1 corresponds to a disaster level of 1), debris flows as DT2 (DT2 corresponds to a disaster level of 2), and industrial fires as DT3 (DT3 corresponds to a disaster level of 3).

[0080] In one embodiment of the present invention, step S2 includes the following steps:

[0081] Specifically, according to the communication stability factor and data redundancy index in the multi-dimensional environmental feature label generated in step S1, the bandwidth adaptation score is calculated by a piecewise linear regression model. The bandwidth adaptation score refers to a numerical evaluation value established based on the quantitative relationship between the real-time change trend of the communication stability factor and the data redundancy index, which is used to characterize the adaptability of the current channel to carry audio and video coding streams. The original audio and video data is subjected to non-local mean noise reduction processing based on wavelet packet thresholds, dynamically matching the time-frequency domain noise energy distribution, and separating the effective signal from the interference component. Non-local mean noise reduction processing refers to the use of the spectrum similarity measurement after wavelet packet decomposition, and the time domain waveform of the audio and video signal is subjected to an energy weight-based noise suppression method.

[0082] The bandwidth adaptation score is calculated using a piecewise linear regression model, satisfying the following formula:

[0083] S band =0.6×(B norm / T avg )+0.4×log10(D level ×C1+1)

[0084] Among them, S band B is the bandwidth adaptation score; norm With T avg The environmental feature label from step S1, B norm is the normalized channel bandwidth value, T avg is the average transmission delay of three consecutive sampling periods; D level is the disaster level; C1 is the audio signal characteristic coefficient; log10 is the logarithmic function with base 10.

[0085] Combined with the satellite communication delay parameters, the dynamic encryption key sequence is generated through Logistic mapping iteration, which satisfies the following formula:

[0086] x n+1 =ρx n (1-x n )+δsin(πt sat )

[0087] Among them, x n 、x n+1 They represent the current value and the next value of the chaotic sequence respectively; ρ represents the branch parameter, which controls the dynamic behavior of the chaotic system; δ is the delay coupling coefficient, which introduces the influence of satellite link delay; t sat The dynamic encryption key sequence is a 32-bit random number block sequence generated by iteratively using the chaotic equation, using the last three millisecond digits of the satellite communication delay as the initial value parameter.

[0088] The spatiotemporal key feature blocks are extracted from the residence interval of the time domain motion vector of the video stream and the audio short-time spectrum. The spatiotemporal key feature blocks refer to the 8×8 pixel blocks in the area where the rate of change of the motion vector between video frames exceeds the preset threshold area, and the spectrum fragments in which the energy of the audio signal after short-time Fourier transform exceeds the background noise by more than 20dB.

[0089] The 4-bit chaotic sequence is embedded in the quantization coefficient residual field of the H.265 coding tree, and a rate-adaptive arithmetic encoder is used to generate a jointly compressed and encrypted bitstream data packet, which satisfies the following formula:

[0090]

[0091] Among them, Q stepis the original quantization step size; K chaos is the chaotic key stream; is the left shift operation.

[0092] The jointly compressed and encrypted code stream data packet refers to the video stream after motion compensation prediction coding and the audio stream after short-time adaptive coding, and the irreversible compressed data stream formed by mixing the encryption factor in the quantization stage.

[0093] For example, in the multi-dimensional environmental feature label generated in step S1, B norm =0.85, T avg =0.38, D level =3, C1 = 2.7, substitute into the above formula to calculate the bandwidth adaptation score S band = 0.6 × (0.85 / 0.38) + 0.4 × log10 (3 × 2.7 + 1) ≈ 1.726. A three-level wavelet packet decomposition was applied to the collected 1080p fire video stream, and a non-local means filter with a noise variance of 0.05 was applied to the 9.6kHz subband. A 256-bit encryption key was generated iteratively, with each 256-bit key consisting of a 128-bit chaotic sequence and a 128-bit delay code.

[0094] Extract the macroblocks with motion vectors exceeding 15 pixels / frame in the flame diffusion area of ​​the video and the continuous high-energy components in the 200-400Hz frequency band of the audio, and convert the 8-bit chaotic key stream K chaos Parity bits of the quantization parameter (QP) embedded in the H.265 coded CU block: HEVC-ENC with an output bit rate of 8Mbps v3a7 Encrypted compressed code stream.

[0095] In one embodiment of the present invention, step S3 includes the following steps:

[0096] Specifically, based on the compressed and encrypted jointly processed code stream output from step S2 and the continuously updated satellite communication environment parameters from step S1, the satellite positioning module obtains the three-dimensional geographic information (longitude, latitude, and elevation) and azimuth data of the disaster site coordinates. A packet loss sensitivity weight matrix is ​​constructed based on the real-time packet loss rate statistics and network jitter distribution characteristics fed back by the satellite link.

[0097] The compressed bitstream is segmented into macroblocks based on keyframe periods. Core layer data is filtered using the contour gradient amplitude of video objects and the baseline slope of the audio spectrum. Enhancement layer texture components are extracted based on the spatial distribution density of motion vector amplitudes. Layered bitrates are dynamically allocated using a regional importance evaluation function. A layered priority index table, updated in real time based on satellite position offsets, is established, forming a dynamic codebook containing basic reconstruction parameters and incremental supplementary data.

[0098] The layered bitrate is dynamically allocated through the regional importance evaluation function, satisfying the following formula:

[0099]

[0100] Among them, W priority is the priority weight coefficient of the hierarchical unit; P loss The packet loss rate of a satellite link is the percentage of the number of packet retransmission requests and the total number of packets sent per unit time, which is the average packet loss rate of the most recent 20 transmission windows. ratio Indicates the degree of change in network transmission delay; T interval is the inverse of the key frame update period; all parameters are dimensionless normalized.

[0101] Satellite positioning information refers to the three-dimensional coordinates of longitude, latitude, and elevation, along with dynamic azimuth parameters, calculated and output by the BD-3 satellite navigation system. The core layer contains the coordinate sequence of I-frame contour control points extracted using the Canny edge detection algorithm, and the fundamental frequency and formant parameter vectors generated by linear predictive coding of the audio signal. A two-level indexing mechanism is used: the first level stores the Canny edge point coordinate sequence (x, y, z), and the second level stores the LPC coefficient matrix. The total data volume is limited to a fixed bandwidth of 3.2Mbps.

[0102] The priority dynamic codebook is a table of stream priority weights calculated based on the geographic coordinate change rate and the packet loss sensitivity matrix. The weight distribution satisfies the constraint that core layer transmission is prioritized in areas with high packet loss. The enhancement layer includes a matrix of DCT coefficient differences for video texture details based on Harris corner detection, an algorithm for detecting image corners using Harris response values, and temporal variation components of motion vectors calculated using optical flow whose modulus exceeds a set threshold.

[0103] For example, when the satellite positioning coordinate offset rate of the petrochemical plant fire scene reaches 0.3 km / h and the packet loss rate continues to be 3.5%, the priority weight coefficient W priority =0.5×0.035+0.3×(0.3 / 0.5)+0.4=0.5975. Canny edge detection is performed on the 15th key frame in the H.265 bitstream to extract 87 contour control points, which are stored as core layer data in the first level.

[0104] Intra-frame texture blocks are calculated for regions where the Harris response exceeds 450, generating 64 8×8 enhancement layer DCT blocks. Weight coefficients are used to allocate a fixed bitrate of 3.2 Mbps for the core layer and a flexible bitrate range of 4.8–6.4 Mbps for the enhancement layer. A DY_CODE_v3 dynamic codebook is constructed, containing the location index [E115°42'N38°23'], the layer identifier CORE_87_CTRL, and ENH_64_DCT. CTRL represents the contour control parameter, and DCT represents the texture transform coefficient.

[0105] In one embodiment of the present invention, step S4 includes the following steps:

[0106] Specifically, the dynamic codebook level data and satellite positioning information generated in step S3 are received through the multi-network fusion evaluation model, and the TCP retransmission rate of the production data network, the MAC layer conflict probability of the broadband ad hoc network, and the frequency statistics of high-throughput satellite beam switching are synchronously obtained. The multi-network fusion evaluation model is a three-dimensional evaluation framework that integrates ground industrial Ethernet communication quality indicators, wireless mesh network topology state parameters, and satellite link switching loss. A network transmission capacity evaluation matrix is ​​constructed based on the real-time congestion index, and the throughput loss cost of the satellite transponder when switching between adjacent beams is calculated. The real-time congestion index is a quantitative value of the network status obtained by weighting the effective bandwidth utilization of the production data network and the backoff delay coefficient of the broadband ad hoc network.

[0107] The dynamic codebook's core and enhancement layer data are used to calculate topology lifetime factors. The Hungarian algorithm is used to match the optimal transmission path combination. Redundant backup channels are established along the satellite link, while data replicas are allocated to critical path nodes in the broadband ad hoc network. Based on the network reliability distribution map, a survivability priority tag is added to each data packet, including a satellite grid code, an upper limit on the number of path switches, and a lifetime timestamp.

[0108] The dynamic codebook layer data refers to the data of the core layer containing the contour control point parameters and the enhancement layer containing the texture differential coefficients divided in step S3. The satellite transponder switching cost refers to the equivalent bandwidth loss caused by the satellite terminal switching between different spot beams. Its value is the product of the switching delay time and the available bandwidth of the current beam.

[0109] Transmission path allocation is the decision-making process of mapping data blocks of different priorities to the optimal transmission medium by establishing a constrained bipartite graph matching model. The indestructible priority tag is a triplet verification identifier that includes a geographic coordinate traceability code, a maximum routing hop limit, and a data validity period.

[0110] The Hungarian algorithm is used to match the optimal transmission path combination, which satisfies the following formula:

[0111]

[0112] Among them, C path represents the path selection cost function; ω1, ω2, ω3 are the weight coefficients of the corresponding parameters; P available is the probability that the redundant path is available.

[0113] E cong is the normalized value of the real-time congestion index, which satisfies the following formula:

[0114]

[0115] Among them, U util Indicates the current bandwidth utilization; U max Indicates the preset bandwidth utilization saturation threshold; τ backoff represents the measured backoff delay; τ norm represents the standard delay benchmark; k represents the delay sensitivity coefficient.

[0116] T handoff is the millisecond-level time loss of satellite beam switching cost, which satisfies the following formula:

[0117] T handoff =Δt switch ×B original ×cosθ

[0118] Where, Δt switch Indicates switching delay; B original represents the original beam width; cosθ represents the angle between beams, which needs to be calculated using ephemeris data.

[0119] For example, the production data network monitors that the effective bandwidth utilization rate reaches 85% (close to the preset saturation threshold of 90%), the self-organizing network backoff delay reaches 12ms (far exceeding the preset normal value of 3ms), and the U util =85% is the current bandwidth utilization, U max =90% is the preset saturation threshold, τ backoff =12ms is the measured backoff delay, τ norm =3ms is the standard delay benchmark, k=0.15 is the delay sensitivity coefficient, and the formula is substituted to obtain:

[0120]

[0121] This index reflects the degree of network congestion. cong ≈0.80 indicates moderate congestion.

[0122] When the satellite link switches from the 115° east longitude beam to the 117° east longitude beam, set Δt switch =220ms is the switching delay, B original=54Mbps is the original beam bandwidth, θ = 15°, and substituting it into the formula yields:

[0123] T handoff =220×54×cos15°≈11476

[0124] This parameter represents the magnitude of the impact of the handover process on service continuity.

[0125] For the CORE_87_CTRL core layer data (including 87 control points) in the DY_CODE_v3 dynamic codebook;

[0126] Preset w1=0.5 as congestion weight, w2=0.3 as switching cost weight, w3=0.2 as path availability weight, T norm =10000 is the standardized reference value, P available =0.9 is the probability of redundant path availability, and the Hungarian algorithm is used to match the optimal transmission path combination:

[0127]

[0128] The calculation result is used for path priority sorting when constructing the dynamic codebook. The lower the value, the better the path quality. The path combination with the lowest cost is selected as the satellite main link + the ad hoc network node 5-7-9 backup path, and the anti-destruction priority tag [BD-E115.7-N38.4 / MAX HOP =3 / TTL=120s], indicating that the satellite grid coordinates are 115.7 degrees east longitude and 38.4 degrees north latitude, with a maximum of 3 hops allowed and a lifetime of 120 seconds.

[0129] In one embodiment of the present invention, step S5 includes the following steps:

[0130] Specifically, the layered codebook data stream marked with the anti-destruction priority tag in step S4 is received, and the coding parameters of the adaptive forward erasure code are selected based on the real-time network topology state. The adaptive forward erasure code dynamically adjusts the redundant coding scheme of the number of coding matrix rows n and the number of original data fragments k based on the network jitter variance, satisfying the constraint that the (k, n) parameter combination adapts to the channel quality. The number of redundant fragments is calculated based on the historical packet loss rate of 10 consecutive transmission cycles:

[0131] R redund =ceil(2.5×P loss ×M)

[0132] Among them, R redund is the number of redundant shards; M is the number of original data shards; P lossis the average packet loss rate of the most recent 10 transmission windows; ceil represents the rounding-up function; and 2.5 is a preset coefficient used to adjust redundancy.

[0133] High-priority data in the core layer is processed using Galois Field generation matrix operations to generate redundant slices. The AES-GCN encryption engine and dynamic bit rate compression module of the satellite portable station hardware chipset are also used for streaming. A dual-threshold decision mechanism is used to divide transmission channels. For urgent data with a lifespan less than a preset threshold in its survivability priority tag, multi-point low-latency transmission links are established through the multi-hop routing table of the broadband ad hoc network. The remaining data streams are interleaved using time division multiplexing (TDM) across idle frequency bands in the high-throughput satellite channel, and the packet header is encapsulated with an index identifier for forward erasure correction codeword information.

[0134] The encryption and compression hardware unit is a dedicated integrated circuit chipset integrated into the satellite portable station, capable of simultaneously executing AES-256 stream encryption and H.265 rate-flexible compression. Redundancy encapsulation is the data processing process of fragmenting the original data and adding checksums to the fragments. This allows the receiver to mathematically recover the original information even if some fragments are lost.

[0135] High-priority data flows refer to codebook core layer information units whose survival timestamp threshold in the indestructibility priority tag is lower than a set value and whose routing hop count is limited. Low-latency ad hoc network links refer to established wireless mesh network routing channels whose end-to-end transmission delay does not exceed a set value. Time-division multiplexing (TDDM) satellite channels are satellite communication modes that divide the channel into periodic time slots and allocate transmission intervals based on priority, while maintaining a constant carrier frequency.

[0136] For example, if the indestructible tag lifetime timestamp of the core layer data in the dynamic codebook is set to 45 seconds and the recent average packet loss rate is 22%, and the number of original data fragments is selected as M=8, the number of redundant fragments R is calculated. redund =ceil(2.5×0.22×8)=ceil(4.4)=5. After the core layer data is divided into 8 original slices, 5 redundant slices are generated by the GF(256) field generation matrix.

[0137] The GF(256) field is a Galois field that supports data redundancy generation and error correction. The hardware unit is called to perform encryption and compression processing at a rate of 150Mbps. For urgent data with a TTL of 45 seconds, the 3-6-9 transmission path of the self-organizing network nodes is allocated to ensure that the data is transmitted within 123ms. Non-critical enhancement layer data is transmitted using 16 time slots of the satellite channel C7 frequency band, and FEC is added to the data packet header. 8+5 The identifier represents the encoding scheme of the original 8 slices + the redundant 5 slices, forming a composite transmission data stream with TSN20240718_1157E timing tags.

[0138] In one embodiment of the present invention, step S6 includes the following steps:

[0139] Specifically, the composite transmission data stream transmitted in step S5 is parsed to extract the satellite grid code and dynamic codebook identifier from the data packet header. Based on the microsecond-level trusted timestamp generated by the satellite timing module carried in the data packet, the iteration threshold generated by the chaotic encryption factor in step S2 is reversed, and the initial key parameters are restored using a dynamic key tree lookup table.

[0140] Among them, the dynamic key tree lookup table organizes key parameters in a tree structure, each node represents a key or key fragment, and the node storing the target key parameters is located in the tree structure based on known key identifiers (such as timestamp, number of iterations and key version number).

[0141] The blockchain node queries the pre-stored smart contract for the satellite grid coordinates, calls the SHA3-256 algorithm, and performs an on-chain verification comparison of the hash digest of the received shards, selecting a subset of shards that meet the consistency check. The dynamic codebook key chain refers to the hash time tree structure constructed based on the chaotic encryption key sequence in step S2, and establishes a key version mapping relationship using the timestamp as the index. The timestamp refers to the absolute time stamp data provided by the satellite RDSS service, a 17-digit precise positioning time code in the format of YYYYMMDDHHMMSSMMM.

[0142] The forward erasure code inverse matrix operation is performed on the verified slices. Combined with the priority dynamic codebook in step S3, the lip mismatch caused by network jitter is eliminated through audio and video synchronization correction, and finally the time-domain aligned fusion streaming data is output.

[0143] Among them, blockchain evidence verification refers to writing the data shard hash value corresponding to the anti-destruction priority tag in step S4 into the blockchain as the transaction content, forming an unalterable distributed verification record.

[0144] The fused audio and video stream refers to a continuous bit stream obtained by resynchronizing and correcting the decoded video inter-frame prediction residual and the audio subband parameters according to the time code in the environmental feature tag in step S1.

[0145] Shard data integrity means that the Hamming distance between the SHA3-256 hash value at the receiving end and the original hash value stored on the blockchain is zero, satisfying the following formula:

[0146]

[0147] Among them, Verif represents the verification result vector; H blockchain Represents the original hash value of the blockchain record; H recv The hash value of the received data shard; It is a bitwise exclusive OR operator. If Verif is all zero, the shard integrity verification is considered to have passed.

[0148] For example, when receiving the composite transmission data stream TSN20240718_1157E, the timestamp 20240718152345000 of the 17-bit precise positioning time code is extracted and parsed as 15:23:45:000 milliseconds on July 18, 2024. The chaotic initial value parameter x0=0.380 in step S2 is traced back to derive the key segment 6A3F generated by the 15345th iteration.

[0149] Query the smart contract address of the 115.7° East Longitude 38.4° North Latitude grid through the blockchain to obtain the shard FEC 8+5 The original hash value is 0x7d3e...c9a2. After calculating the local hash of the 13 received slices (8 original + 5 redundant), the verification finds that the calculated value of the 7th slice, 0x7d3e...c9b4, does not match the original hash value 0x7d3e...c9a2. If Verif = 0x0012 is not all zeros, the verification fails, triggering the redundant slice R3 to reconstruct the failed 7th slice. Finally, the decoded contour control point coordinates are used to reconstruct the video keyframes, synchronized with the Mel-spectrogram parameters, and output a fused stream of 1920×1080@60fps video and 48kHz audio.

[0150] See attached Figure 2 The present invention also proposes a satellite portable station audio and video data compression system, comprising the following modules:

[0151] The data acquisition module is used to collect raw audio and video data, synchronously analyze the current satellite communication environment parameters and disaster types, and generate multi-dimensional environmental feature labels;

[0152] The pre-processing module is used to score the bandwidth of environmental feature tags, adaptively reduce noise on raw audio and video data, generate dynamic encryption key sequences, and generate compressed and encrypted jointly processed bitstreams;

[0153] The dynamic codebook partitioning module is used to extract the data of the core layer and the enhancement layer in layers according to the packet loss rate and satellite positioning information to form a priority dynamic codebook;

[0154] Path allocation module, used to allocate transmission paths and add anti-destruction priority tags through the multi-network fusion evaluation model;

[0155] Redundant encapsulation module, which uses adaptive forward erasure coding to generate redundant fragments, directing high-priority data streams to low-latency ad hoc network links, and multiplexing non-critical data streams into high-throughput satellite channels;

[0156] The verification and reconstruction module is used to trace back the key chain and verify the integrity of the fragments, and generate a fused stream through audio and video synchronization.

[0157] It should be noted that the formulas described above can translate physical quantities of different attributes into unitless standard values ​​or superimposable parameters of the same dimension through the principle of dimensional consistency and mathematical standardization (e.g., normalization, dimensionless parameter conversion, or unit system unification). This eliminates the interference of different dimensions on the operational logic, allowing the formulas to retain the distribution characteristics of the original data while maintaining mathematical rationality and adaptability to objective laws. The above are merely exemplary embodiments of the present invention and are not intended to limit the scope of the present invention.

[0158] The modules can be implemented in whole or in part through software, hardware, or a combination thereof, supporting hardware embedded in or independent of a processor in a computer device, and also supporting software stored in a memory in a computer device, so that the processor can call and execute operations corresponding to the modules.

[0159] It should be noted that the human body information (including but not limited to human device information and personal information, etc.) and data (including but not limited to data used for analysis, stored data and displayed data, etc.) involved in the present invention are all information and data authorized by the human body or fully authorized by all parties. The collection, use and processing of relevant data require relevant legal standards.

[0160] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. A satellite portable station audio and video data compression method, characterized in that: The following steps are involved: S1. Obtain the original audio and video data collected by the emergency site sensors, synchronously analyze the current satellite communication environment parameters and disaster types, and generate multi-dimensional environmental feature labels; S2: Based on the bandwidth score of the environmental feature label, adaptive noise reduction is performed on the original audio and video data, and a dynamic encryption key sequence is generated. The key spatiotemporal features are extracted and embedded into the chaotic encryption factor to generate a compressed and encrypted joint processing stream. S3. Dynamically divide the compressed primary code stream into layers based on the real-time satellite link packet loss rate and Beidou positioning information: the core layer contains key frame outlines and spectrum baseline parameters, and the enhancement layer contains texture details and motion vector data, forming a priority dynamic codebook; S4. Utilize a multi-network convergence assessment model, based on the real-time congestion index of the production data network and broadband ad hoc network, and the switching cost of high-throughput satellite transponders, to allocate transmission paths for each layer of the dynamic codebook and assign anti-destruction priority tags to critical data. S5. Redundancy encapsulation of the fragmented data set is performed through adaptive forward erasure coding. Combined with the encryption and compression hardware unit built into the satellite portable station, high-priority data streams are directed to the low-latency ad hoc network link, while non-critical data streams are multiplexed on the high-throughput satellite channel. S6. The receiving end uses the timestamp to trace back the key chain of the dynamic codebook, and verifies the integrity of each fragment data through blockchain evidence to reconstruct the original semantic fusion audio and video stream.

2. A satellite portable station audio and video data compression method according to claim 1, characterized in that: Generating multi-dimensional environmental feature labels includes the following steps: The data acquisition terminal composed of high-sensitivity cameras, microphone arrays, and temperature sensors deployed at the emergency site collects the original audio and video data of the scene in real time; Obtain the signal transmission delay, transmission bandwidth, carrier-to-noise ratio, and channel bit error rate parameters of the current satellite communication channel, and construct a characteristic sequence of real-time communication parameters through fast Fourier transform; Conduct joint analysis of infrared thermal imaging data and terrain vibration spectrum data collected at the disaster site, and use a pre-trained disaster classification model to identify the disaster type; The time-frequency characteristics of the original audio and video data, the characteristic sequence of real-time communication parameters and the disaster type are multi-dimensionally weighted fused to generate a multi-dimensional environmental feature label including communication stability factor, disaster level and data redundancy index.

3. A satellite portable station audio and video data compression method according to claim 2, characterized in that: The time-frequency characteristics of the original audio and video data, the characteristic sequence of real-time communication parameters and the disaster type are multi-dimensionally weighted fused, including the following steps: The raw audio and video data is collected by a sensor module consisting of a wide-angle high-definition camera and a circular microphone array at the disaster site, continuously collecting uncompressed sound wave signals and optical image sequences; The real-time communication environment parameters are a set of communication channel physical layer parameters measured by an on-site satellite modem, including a measured value of instantaneous available bandwidth, a signal round-trip delay, and a linear quantized value of a carrier-to-noise power ratio; The disaster type, which is the recognition result generated by the trained convolutional neural network based on the input on-site infrared spectrum image and the frequency domain characteristics of the foundation micro-vibration, includes preset disaster types such as earthquakes, debris flows, and industrial fires; Based on the normalized feature vectors established by three types of indicators, namely audio and video information capacity, communication channel availability and disaster hazard degree, the three types of indicators are mapped into a transmittable label coding group through a weighting factor matrix.

4. A satellite portable station audio and video data compression method according to claim 3, characterized in that: Generating a dynamic encryption key sequence includes the following steps: Based on the last three digits of the millisecond level of satellite communication delay as the initial value parameter, a 32-bit random number grouping sequence is generated through the iteration of the Logistic chaos equation; The iteratively generated chaotic sequence is bit-wise concatenated with the satellite communication delay parameter to generate a dynamic encryption key sequence; The chaotic sequence is embedded in the quantization coefficient residual domain of the H.265 coding tree, and a rate-adaptive arithmetic encoder is used to generate jointly compressed and encrypted bitstream data packets.

5. A satellite portable station audio and video data compression method according to claim 1, characterized in that: Dynamic hierarchical partitioning includes the following steps: The core layer data is filtered using the contour gradient amplitude of the video object and the baseline slope of the audio spectrum; Extracting the enhancement layer texture component according to the spatial distribution density of the motion vector amplitude; The layered code rate is dynamically allocated through the regional importance evaluation function, and a layered priority index table is established which is corrected in real time by the satellite position offset, forming a dynamic codebook containing basic reconstruction parameters and incremental supplementary data.

6. A satellite portable station audio and video data compression method according to claim 1, characterized in that: Allocating a transmission path includes the following steps: Obtain statistics on the TCP retransmission rate of the production data network, the MAC layer conflict probability of the broadband ad hoc network, and the frequency of high-throughput satellite beam switching; A network transmission capacity evaluation matrix is ​​constructed based on the real-time congestion index to calculate the throughput loss cost of satellite transponders when switching between adjacent beams. The topology lifetime factors of the core layer and enhancement layer data of the dynamic codebook are calculated respectively. The Hungarian algorithm is used to match the optimal transmission path combination, and an anti-destruction priority tag including the satellite grid code, the upper limit of the path switching number and the survival time stamp is added.

7. A satellite portable station audio and video data compression method according to claim 1, characterized in that: Performing redundancy encapsulation includes the following steps: Select encoding parameters of adaptive forward erasure code based on real-time network topology status; The core layer uses Galois Field generation matrix operations to generate redundant slices for high-priority data; For urgent data with a lifetime less than a preset threshold in the indestructibility priority tag, a multi-point low-latency transmission link is established through the multi-hop routing table of the broadband ad hoc network.

8. A satellite portable station audio and video data compression method according to claim 1, characterized in that: Blockchain evidence verification includes the following steps: The iteration threshold of the chaotic encryption factor is inferred based on the satellite timing timestamp, and the initial key parameters are restored through the dynamic key tree lookup table. The blockchain node queries the pre-stored smart contract under the satellite grid coordinates, and calls the SHA3-256 algorithm to perform on-chain evidence traversal and comparison on the hash summary of the received shard; Filter the subset of shards that pass the consistency check to trigger the reconstruction of redundant shards.

9. A satellite portable station audio and video data compression method according to claim 8, characterized in that: Restore the initial key parameters through the dynamic key tree lookup table, including: Each node represents a key or key fragment. Based on the known key identifier containing the timestamp, iteration number and key version number, the node storing the target key parameters is located in the tree structure.

10. A satellite portable station audio and video data compression system, characterized in that: Includes the following modules: The data acquisition module is used to collect raw audio and video data, synchronously analyze the current satellite communication environment parameters and disaster types, and generate multi-dimensional environmental feature labels; The pre-processing module is used to score the bandwidth of environmental feature tags, adaptively reduce noise on raw audio and video data, generate dynamic encryption key sequences, and generate compressed and encrypted jointly processed bitstreams; The dynamic codebook partitioning module is used to extract the data of the core layer and the enhancement layer in layers according to the packet loss rate and satellite positioning information to form a priority dynamic codebook; Path allocation module, used to allocate transmission paths and add anti-destruction priority tags through the multi-network fusion evaluation model; Redundant encapsulation module, which uses adaptive forward erasure coding to generate redundant fragments, directing high-priority data streams to low-latency ad hoc network links, and multiplexing non-critical data streams into high-throughput satellite channels; The verification and reconstruction module is used to trace back the key chain and verify the integrity of the fragments, and generate a fused stream through audio and video synchronization.

Images