A load original code stream fast screening method based on track extrapolation model

Abstract

This invention relates to the field of satellite remote sensing data processing technology, and in particular to a method for rapid screening of raw payload bitstreams based on an orbit extrapolation model. The method includes the following steps: a ground system receives the raw payload bitstream from the downlink satellite in real time; based on a pre-constructed imaging geometry model and the extrapolated precise orbit information of the imaging arc segment, the starting position of the raw bitstream data containing the target geographic region within the raw payload bitstream is predicted; using data packets as processing units, starting from the starting position, multiple data packets are processed in parallel with decryption, decompression, and raw bitstream data segment selection to extract the target bitstream segment corresponding to the target geographic region; based on the target bitstream segment, a level 0 slice product is generated and output. This invention improves the preprocessing efficiency of the ground receiving system, shortens the cycle from remote sensing data acquisition to application, and reduces the burden on subsequent processing through accurate target data extraction.

Description

Technical Field

[0001] This invention relates to the field of satellite remote sensing data processing technology, and in particular to a method for rapid screening of raw payload bitstreams based on an orbit extrapolation model. Background Technology

[0002] We live in the space age, and the application of satellite technology has had a wide-ranging impact on people's production and lives. Among them, the application of satellite remote sensing technology is the most widespread. From resource surveys, environmental monitoring, disaster emergency response, agricultural production, urban planning and management, meteorological and oceanographic monitoring, to national defense and security needs, everyone has high expectations for the effective payload data of remote sensing satellites.

[0003] Existing remote sensing raw data stream processing technologies mainly revolve around the process of "on-board acquisition and preprocessing - space transmission - ground reception and post-processing". Core technologies include data compression, data encryption, data framing and coding, modulation and demodulation, and ground preprocessing (such as radiometric calibration and atmospheric correction). Among them, data compression mostly adopts traditional lossy or lossless compression algorithms, the transmission stage mainly adopts X-band and Ka-band radio frequency transmission combined with variable coding modulation technology, and ground processing relies on offline batch processing mode to complete data stream parsing and data calibration.

[0004] However, existing technical solutions have the following shortcomings in practical applications: Processing efficiency is not matched with data growth: As remote sensing satellites enter the era of "three highs" (high spatial, high temporal, and high spectral resolution), the amount of raw bitstream data is growing exponentially. The existing ground batch processing mode is time-consuming. According to statistics, about 80% of the time and effort is devoted to the data preparation stage, resulting in an excessively long cycle from data acquisition to usability, which makes it difficult to meet the needs of rapid application. Weak satellite-ground collaborative processing capability: The satellite's processing capability is limited, and it can only complete basic frame coding. A large number of complex processing tasks (such as radiometric correction, geometric correction, target extraction, etc.) are forced to be processed on the ground. When the transmission bandwidth is insufficient, a large amount of data needs to be stored on the satellite for a long time, which not only increases the storage pressure, but also reduces the utilization efficiency and timeliness of the payload. Poor compatibility of multi-source streams: The original stream formats of different satellite platforms and payloads vary greatly; the existing processing system lacks a unified adaptation mechanism, and the preprocessing complexity is high when multi-source data is processed collaboratively, which can easily lead to inconsistent analysis results and increase the difficulty of system integration. Insufficient real-time processing capabilities: Traditional processing architectures employ linear and serial processing flows, which are insufficient to meet the real-time requirements for rapid data analysis and application in scenarios such as disaster emergency response and military reconnaissance, and thus cannot provide timely and effective support for decision-making.

[0005] Therefore, how to quickly and accurately filter out target area data of interest to users from massive, high-speed satellite raw bitstreams, and efficiently complete preprocessing, has become a pressing technical problem to be solved in the field of remote sensing satellite data processing. Summary of the Invention

[0006] In order to overcome the problems of low processing efficiency, poor real-time performance, and difficulty in quickly locating target area data in existing remote sensing raw bitstream processing technologies.

[0007] The technical solution of this invention is: a method for fast screening of raw load bitstreams based on an orbit extrapolation model, comprising the following steps: Data stream reception steps: The ground system receives the raw bit stream of the downlink payload from the satellite in real time; Target bitstream location prediction step: Based on the pre-built imaging geometry model and the extrapolated imaging arc precise trajectory information, predict the starting position of the original bitstream data containing the target geographical region in the original bitstream of the payload; Parallel streaming processing steps: Using data packets as processing units, starting from the starting position, decryption processing, decompression processing, and original bitstream data segment selection processing are performed in parallel on multiple data packets to extract the target bitstream segment corresponding to the target geographical region; Product output steps: Based on the target bitstream segment, generate and output the level 0 slice product.

[0008] Preferably, this scheme introduces an orbit extrapolation model and an imaging geometry model to predict the location of the target data before or in the early stages of receiving the bitstream, thereby avoiding serial processing of the entire data. At the same time, the parallel streaming processing architecture greatly improves processing efficiency. Through the above steps, the system can start processing directly from the predicted starting position, and only decrypt, decompress, and filter the data containing the target area, reducing the waste of computing resources and the scale of data processing.

[0009] Preferably, the target bitstream location prediction step further includes: Construct the imaging geometric model that includes satellite orbital parameters, attitude parameters, and sensor imaging parameters; Based on the satellite's predetermined mission plan or real-time received auxiliary data, the precise orbital information covering the imaging time period of the target geographical area is extrapolated. By combining the imaging geometry model and the precise orbit information, the corresponding pixel position of the target geographic area on the satellite imaging focal plane is calculated, and then the frame number range or timestamp range of the original bitstream containing the pixel position data in the whole orbit data is predicted to determine the starting position.

[0010] Preferably, in the parallel streaming processing step, the data packet consists of a fixed number of data frames, and the decryption processing, decompression processing, and original code stream data segment selection processing are executed in parallel on multiple processing units in a pipeline manner.

[0011] Preferably, the original bitstream data segment selection process in the parallel streaming processing step further includes: Analyze the auxiliary data in the current processing unit to obtain real-time positioning information; Based on the real-time positioning information, update the imaging geometric model or orbit extrapolation result, and determine whether the imaging area corresponding to the current processing unit intersects with the target geographical area; If an intersection exists, the image data portion in that processing unit is marked as part of the target bitstream segment.

[0012] Preferably, determining whether the imaging area corresponding to the current processing unit intersects with the target geographical area specifically includes: For point targets, the vector product method is used to determine whether the point target is located within the quadrilateral formed by the latitude and longitude coordinates of the four corner points of the imaging area corresponding to the current processing unit; For a target area, the vector product method is used to determine whether the boundary points of the target area fall within the quadrilateral formed by the latitude and longitude coordinates of the four corner points of the imaging area corresponding to the current processing unit.

[0013] Preferably, the decryption process includes: identifying and removing invalid frames, decrypting valid data according to preset decryption rules and a key library, and removing the encryption header.

[0014] Preferably, the decompression process includes: decompressing the decrypted data according to a preset decompression algorithm, and separating the decompressed data stream into auxiliary data and image data.

[0015] Preferably, in the product output step, the level 0 slice product includes the target bitstream segment and its corresponding auxiliary data.

[0016] Preferably, the method further includes a preprocessing step: prior to the parallel streaming processing step, preloading the radiometric correction coefficients and / or reference images corresponding to the target geographic area for radiometric correction of the subsequently generated Level 0 tile products.

[0017] Preferably, the method further includes a geometric correction step: using some satellite parameters that are resolved in real time in the parallel streaming processing step, the system performs geometric correction on the Level 0 slice product by using a local parameter to predict the global value, and generates a Level 1 slice product.

[0018] The beneficial effects of this invention are: 1. This invention adopts a prediction-localization-parallel streaming processing architecture. Through orbit extrapolation and imaging geometry models, the starting position of the target data in the bitstream can be predicted at the beginning of data reception or even before reception, thus avoiding the processing of massive amounts of irrelevant data. The parallel pipeline processing architecture based on data packets enables data to be received, decrypted, decompressed, and filtered simultaneously. This mechanism shortens the delay from data reception to usability (generation of level 0 slices), allowing the delay of extracting data from a target area of ​​5km×5km from the original bitstream and generating level 0 products to be controlled within seconds, improving processing efficiency and real-time performance. 2. This invention combines extrapolated precise trajectory information with an imaging geometric model to calculate the precise location of the target geographical region in the original bitstream. During processing, secondary verification is performed using auxiliary data obtained from real-time parsing, and precise algorithms such as the vector product method are employed to determine the inclusion relationship between the imaging region and the target region. This ensures that the extracted target bitstream segment is both complete and minimized, containing only the region data that the user is truly interested in. This not only reduces the pressure on data storage and transmission but also reduces the input data scale for subsequent radiometric correction, geometric correction, and other processing steps, further improving the processing efficiency of the entire chain. It achieves accurate target data positioning and extraction, reducing data redundancy. 3. The parallel streaming processing framework of this invention has good modularity. Core functions such as decryption, decompression, and data segment selection are all implemented in a modular form. By preloading different decryption libraries and decompression algorithm libraries, it can quickly adapt to the original bitstream formats of different satellite platforms and payloads, solving the problems of poor compatibility and high preprocessing complexity of existing multi-source bitstreams. At the same time, the framework supports the flexible insertion of other processing modules into the process, such as radiometric correction preprocessing and fast geometric correction, providing a technical foundation for building a unified, efficient, and scalable ground data processing system and enhancing the system's flexibility and compatibility. Attached Figure Description

[0019] Figure 1 The diagram shows the technical roadmap of the load raw code stream fast screening method based on the orbit extrapolation model of the present invention. Figure 2 The diagram shown is a production timing diagram of the original code stream fast screening method based on the track extrapolation model of the present invention. Figure 3 The diagram shown illustrates the principle of selecting raw code stream data segments in the load raw code stream fast screening method based on the track extrapolation model of the present invention. Figure 4 The diagram shows the production process of the target area raw code stream fast screening method based on the track extrapolation model of the present invention. Detailed Implementation

[0020] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0021] Example 1: Please see Figure 1 , Figure 2 , Figure 3 and Figure 4 This invention provides an embodiment: a method for rapid screening of raw load bitstreams based on an orbit extrapolation model, comprising the following steps: Data stream reception steps: The ground system receives the raw bit stream of the downlink payload from the satellite in real time; Target bitstream location prediction step: Based on the pre-built imaging geometry model and the extrapolated imaging arc precise trajectory information, predict the starting position of the original bitstream data containing the target geographical region in the payload original bitstream; Parallel streaming processing steps: Using data packets as processing units, starting from the beginning position, multiple data packets are processed in parallel to perform decryption, decompression, and raw bitstream data segment selection to extract the target bitstream segment corresponding to the target geographical region; Product output steps: Based on the target bitstream segment, generate and output the level 0 slice product.

[0022] This scheme introduces an orbit extrapolation model and an imaging geometry model to predict the location of target data before or in the early stages of receiving the bitstream, thus avoiding serial processing of the entire data. At the same time, the parallel streaming processing architecture greatly improves processing efficiency. Through the above steps, the system can start processing directly from the predicted starting position, decrypting, decompressing, and filtering only the data containing the target area, reducing the waste of computing resources and the scale of data processing.

[0023] Furthermore, the target bitstream location prediction step further includes: Construct an imaging geometric model that includes satellite orbital parameters, attitude parameters, and sensor imaging parameters; Based on the satellite's predetermined mission plan or real-time received auxiliary data, the precise orbital information for the imaging time period covering the target geographical area is extrapolated. By combining the imaging geometry model and precise orbit information, the corresponding pixel position of the target geographic area on the satellite imaging focal plane is calculated, and then the frame sequence number range or timestamp range of the original bitstream containing the pixel position data in the whole orbit data is predicted to determine the starting position. By constructing a geometric model and extrapolating the trajectory, the theoretical position of the target area on the image plane can be calculated, thereby inversely deducing the original bitstream frame number or timestamp corresponding to the theoretical position. This prediction method based on a physical model can obtain target position information earlier than simply relying on auxiliary data for real-time judgment.

[0024] Furthermore, in the parallel streaming processing steps, the data packets consist of a fixed number of data frames. Decryption, decompression, and raw bitstream data segment selection are executed in parallel on multiple processing units in a pipeline manner. By dividing the data stream into fixed-size data packets and adopting a pipelined parallel processing mode, operations such as decryption, decompression, and selection can be performed simultaneously. For example, while processing unit 1 is decrypting data packet A, processing unit 2 can simultaneously decompress data packet B. This mechanism effectively hides the delays of each processing stage, enabling data to be received, processed, and output simultaneously, thus shortening the time for the first Level 0 product to be output.

[0025] Furthermore, the raw bitstream data segment selection process in the parallel streaming processing step further includes: Analyze the auxiliary data in the current processing unit to obtain real-time positioning information; Based on real-time positioning information, update the imaging geometric model or orbit extrapolation results, and determine whether the imaging area corresponding to the current processing unit intersects with the target geographic area; If an intersection exists, the image data portion in that processing unit is marked as part of the target bitstream segment; The scheme describes the logic for making accurate judgments during parallel processing. Although there is a pre-predicted starting position, due to errors in orbit extrapolation or uncertainties in the target area boundary, it is still necessary to use real-time parsed auxiliary data (such as the latest GPS positioning information) for secondary confirmation during processing. By judging the spatial relationship between the imaging area and the target area, the integrity and accuracy of the target code stream segment extraction are ensured.

[0026] Furthermore, determining whether the imaging area corresponding to the current processing unit intersects with the target geographical area specifically includes: For point targets, the vector product method is used to determine whether the point target is located within the quadrilateral formed by the latitude and longitude coordinates of the four corner points of the imaging area corresponding to the current processing unit; For regional targets, the vector product method is used to determine whether the boundary points of the regional targets fall within the quadrilateral formed by the latitude and longitude coordinates of the four corner points of the imaging area corresponding to the current processing unit; This scheme provides a specific algorithm for spatial intersection determination; the vector product calculation method is an efficient algorithm for determining whether a point is inside a polygon, which is suitable for real-time processing scenarios; by simplifying the imaging area into a quadrilateral and the target into a point or multiple boundary points, this algorithm can be used to quickly and accurately determine spatial relationships, providing a reliable mathematical basis for data segment extraction.

[0027] Furthermore, the decryption process includes: identifying and removing invalid frames, decrypting valid data according to preset decryption rules and a key library, and removing the encryption header. Among these steps, since errors or frame loss may occur during spatial link transmission, removing invalid frames first ensures the data quality of subsequent processing. Decryption according to preset rules and a key library ensures data security. Removing the encryption header is to restore the clean payload data stream, preparing for subsequent decompression and parsing.

[0028] Furthermore, the decompression process includes: decompressing the decrypted data according to a preset decompression algorithm, and separating the decompressed data stream into auxiliary data and image data; different satellite payloads may use different compression algorithms (such as JPEG-LS, CCSDS recommended standards, etc.), so decompression according to a preset algorithm is required; after decompression, the mixed data stream is separated into structured auxiliary data (such as orbit, attitude, timecode, etc.) and raw image data, which is the basis for subsequent positioning judgment and product generation.

[0029] Furthermore, in the product output step, the Level 0 slice product includes the target bitstream segment and its corresponding auxiliary data; the output product not only includes the selected image data, but also must include the auxiliary data that is strictly synchronized with it; this ensures the data independence and integrity of the Level 0 slice product, making it a standardized input for subsequent advanced processing such as radiometric correction and geometric correction.

[0030] Furthermore, the method also includes a preprocessing step: before the parallel streaming processing step, the radiometric correction coefficients and / or reference images corresponding to the target geographic area are preloaded for radiometric correction of the subsequently generated Level 0 tile products; wherein, radiometric correction usually requires looking up the corresponding correction coefficient library; by preloading the correlation coefficients of the target area into memory before the processing starts, frequent I / O operations during the processing can be avoided, so that the radiometric correction process can be started immediately once the Level 0 tile product is generated, further compressing the total time from the raw bitstream to the usable product.

[0031] Furthermore, the method also includes a geometric correction step: using some satellite parameters parsed in real time during the parallel streaming processing step, a local parameter prediction method is used to perform systematic geometric correction on the Level 0 tile product to generate a Level 1 tile product. Since traditional geometric correction requires auxiliary parameters of the entire scene data, this solution uses some parameters that have been parsed in the parallel streaming processing and cover the target tile to predict the geometric correction model required for the entire tile through interpolation or extrapolation, thereby achieving rapid geometric correction of the Level 0 tile and generating a Level 1 tile product with georeferenced information, which meets the need for rapid positioning information in emergency scenarios.

[0032] Through the above steps, this invention adopts a prediction-localization-parallel streaming processing architecture. By using an orbit extrapolation model and an imaging geometry model, the starting position of the target data in the bitstream can be predicted at the beginning of data reception, or even before reception, thus avoiding the processing of massive amounts of irrelevant data. A parallel pipeline processing architecture based on data packets is adopted, enabling simultaneous data reception, decryption, decompression, and filtering. This mechanism shortens the delay from data reception to usability (generating level 0 slices), allowing the delay of extracting data from a 5km×5km target area from the original bitstream and generating a level 0 product to be controlled within seconds, improving processing efficiency and real-time performance. This invention, by combining the precise extrapolated orbit information with the imaging geometry model, can calculate the precise location of the target geographical area in the original bitstream. During processing, secondary verification is performed using auxiliary data obtained from real-time parsing, and precise algorithms such as the vector product calculation method are used to determine the inclusion relationship between the imaging area and the target area. This ensures that the extracted target bitstream segments are both complete and minimized, containing only data from areas of genuine interest to the user. This not only reduces the pressure on data storage and transmission but also reduces the input data size for subsequent radiometric and geometric correction processes, further improving the overall processing efficiency. It achieves accurate target data location and extraction, reducing data redundancy. The parallel streaming processing framework of this invention has excellent modularity; core functions such as decryption, decompression, and data segment selection are all implemented in a modular form. By pre-loading different decryption libraries and decompression algorithm libraries, it can quickly adapt to the original bitstream formats of different satellite platforms and payloads, solving the problems of poor compatibility and high preprocessing complexity of existing multi-source bitstreams. Simultaneously, the framework supports the flexible insertion of other processing modules into the process, such as radiometric correction preprocessing and rapid geometric correction, providing a technical foundation for building a unified, efficient, and scalable ground data processing system, enhancing the system's flexibility and compatibility.

[0033] Example 2: Optionally, this embodiment provides a method for rapid screening of raw payload streams based on an orbit extrapolation model. The method described in this embodiment is applied to a satellite ground receiving station to process the raw streams transmitted from a high-resolution optical remote sensing satellite. The goal is to quickly screen out image data covering a specific geographical area (e.g., the range of 116.0°-116.5°E and 39.5°-40.0°N) from the massive streams and generate level 0 tile products.

[0034] The method includes the following specific steps: Step 1: Data stream reception: The ground system initiates the data receiving procedure, tracks the satellite through the antenna, and receives the raw bitstream of the downlink payload from the satellite in real time; this bitstream is a binary data stream transmitted continuously in units of fixed-length data frames (e.g., each frame contains 1024 bytes).

[0035] Step 2: Target bitstream position prediction: Before or in the early stages of data reception, a position prediction operation is performed to determine the starting position of the target data within the entire track of data to be received. This step specifically includes: (1) Constructing an imaging geometric model: A rigorous imaging geometric model is pre-established based on the satellite's orbital parameters (such as semi-major axis a, eccentricity e, inclination i), attitude parameters (such as roll angle φ, pitch angle θ, yaw angle ψ), and sensor internal parameters (such as focal length f, pixel size d, and sensor arrangement); this model can be represented as a mapping relationship from ground point coordinates (X, Y, Z) to image point coordinates (s, l): Where [X_s(t), Y_s(t), Z_s(t)]^T is the position vector of the satellite in the spatial rectangular coordinate system at time t, R(t) is the rotation matrix determined by the satellite attitude at time t, [x, y, -f]^T is the coordinate of the image point in the image plane coordinate system, and λ is the scaling factor.

[0036] (2) Extrapolation of precise orbit: Based on the precise orbital elements of the satellite the previous day or several orbits, the precise position and velocity vector of the satellite during the imaging time window (expected to be from 12:00:00 to 12:00:20 UTC) of the target area during this transit is extrapolated using the SGP4 / SDP4 simplified gravity model or a more precise numerical integration method, which is to obtain [X_s(t), Y_s(t), Z_s(t)]^T in step (1).

[0037] (3) Calculate the bitstream position corresponding to the target location: Substitute the geographic coordinates (latitude and longitude) of the four corner points of the target area into the inverse calculation process of the above imaging geometry model to calculate the theoretical pixel coordinates (row number l, column number s) of these ground points on the satellite image; row number l corresponds to the imaging time; from this, the imaging time t_target corresponding to the center point of the target area can be calculated; then, according to the correspondence between the time code of the satellite platform and the data frame count (for example, 1000 frames are downloaded per second, and each frame corresponds to one row of scanning), t_target is converted into the starting value of the frame count Frame_start; for example, if the imaging start time is 12:00:00, which corresponds to a frame count of 0, then Frame_start = (t_target - 12:00:00) × 1000 - Δ, where Δ is the frame number offset reserved to ensure complete coverage of the target area.

[0038] Step 3: Parallel streaming processing: The ground processing system initiates parallel streaming processing on a packet-by-packet basis; packets consist of a fixed number of data frames (e.g., 128 frames); starting with the packet corresponding to the predicted starting frame count Frame_start, subsequent packets are fed into the parallel processing pipeline; the system comprises three main processing stages, each executed in parallel in a pipelined manner on multiple processing cores: (1) Decryption processing: Processing unit A is responsible for decryption; first, it identifies and removes invalid frames containing error flags; then, according to the decryption rules agreed upon in advance by the satellite mission (such as the AES-256 algorithm) and the key library stored in the local hardware security module, it decrypts the valid data frames and removes the encryption identifier field in the frame header.

[0039] (2) Decompression processing: Processing unit B is responsible for decompression; receiving the decrypted data packets and removing invalid frames (if any) found during the decryption process; decompressing the image data part according to the compression algorithm used by the satellite payload (e.g., the wavelet-based image compression algorithm recommended by the CCSDS 122.0-B-1 standard) to recover the original pixel values; the decompressed data stream is separated into two parts: one is a structured auxiliary data packet (containing more detailed GPS time, attitude quaternions, etc.), and the other is the image data block.

[0040] (3) Original bitstream data segment selection and processing: Processing unit C is responsible for selection; parsing the auxiliary data separated by the decompression step to obtain the precise imaging time t_current and satellite position corresponding to the current data packet; comparing the position at this moment with the pre-loaded target area polygon to determine whether the current data packet actually contains target area data; the judgment method adopts the vector product calculation method detailed in Example 3 below; if it is determined to contain, the image data block in the processing unit along with its auxiliary data is marked and cached as part of the target bitstream segment.

[0041] The three processing stages described above perform different data packets simultaneously. For example, while processing core 1 decrypts data packet N, processing core 2 simultaneously decompresses data packet N-1, which has already been decrypted by core 1, and processing core 3 selects data segments from data packet N-2, which has been decompressed by core 2. This pipeline design significantly hides processing latency.

[0042] Step 4: Product Output Once the current data packet is selected and determined to be the last data packet constituting the target area, the system integrates all the marked and cached target bitstream segments (i.e., image data blocks and their corresponding auxiliary data within multiple data packets) and encapsulates them into an independent data file, namely the Level 0 slice product. The format of this product file can follow a standard format (such as the original band of GeoTIFF or HDF5), and the auxiliary data contained therein ensures the independence of the slice.

[0043] Through the above steps, this embodiment achieves rapid prediction, location and extraction of the original code stream of the target area, avoiding the processing of full-track data and significantly improving processing efficiency.

[0044] Example 3: Optionally, this embodiment further refines the spatial intersection judgment method in the original code stream data segment selection process based on embodiment 2.

[0045] In this embodiment, the "raw bitstream data segment selection processing" in step 3 uses the following specific algorithm to determine whether the imaging area corresponding to the current data packet intersects with the target geographical area: First, the auxiliary data in the current processing unit is parsed to obtain the precise imaging times t_start and t_end corresponding to the start and end frames of the data packet. Combined with the pre-loaded precise orbit data and imaging geometry model, the satellite position and attitude corresponding to these two frames can be calculated. Then, using the collinearity equation, the four corner points of the imaging sensor are projected onto the Earth's surface to obtain the quadrilateral range of the imaging area corresponding to the current data packet on the Earth's surface. The latitude and longitude coordinates of the four corner points are P1(Lon1, Lat1), P2(Lon2, Lat2), P3(Lon3, Lat3), and P4(Lon4, Lat4). This quadrilateral approximately represents the ground area covered by the current data packet.

[0046] The target area can be a point or a surface area: Scenario A: Targeting point objects (such as a single building or hotspot) The geographic coordinates of the target point are set as T(Lon_t, Lat_t); to determine whether point T is inside quadrilateral P1P2P3P4, the vector product method is used; the principle is: if a point is inside a convex polygon, then the sum of the areas of the triangles formed by the point and each side of the polygon equals the area of ​​the polygon itself; or by determining whether the point is always located on the same side of each edge; the specific algorithm is as follows: For quadrilateral P1P2P3P4, calculate the vector product: For two-dimensional vectors a=(x1, y1) and b=(x2, y2), the z-component of their vector product is defined as x1×y2 - y1×x2; the sign of this value indicates whether the point is located on the left or right side of the directed line segment; if V1, V2, V3, and V4 are all greater than or equal to 0 or all less than or equal to 0, then point T is determined to be inside (or on) the polygon P1P2P3P4; at this time, the inclusion flag is set to 1, indicating that the current data packet contains the target data.

[0047] Scenario B: Targeting regional objectives (such as administrative regions or lakes) The target region is defined by its boundary point sequence Q1, Q2, ..., Qm (where m is the number of boundary points); to determine whether the region intersects with quadrilateral P1P2P3P4, a traversal method is used: (1) Traverse each boundary point Qi of the target region; (2) For each boundary point Qi, use the same vector product calculation method as in case A to determine whether Qi is inside quadrilateral P1P2P3P4; (3) If any boundary point Qi is inside the quadrilateral, it is immediately determined that there is an intersection, the inclusion flag is set to 1, and the current and subsequent data packets are extracted until the flag becomes 0; (4) If no point is found inside the quadrilateral after traversing all boundary points, but the polygon may completely contain the quadrilateral, or the quadrilateral may completely contain the polygon, further judgment is needed; as a simplified and efficient method, the center point C of the quadrilateral can be calculated and it can be determined whether C is inside the target polygon. If C is inside the target polygon, it is also determined that there is an intersection, and flag=1; if it is still not, it is determined that there is no intersection, and flag=0.

[0048] Based on the value of flag, the system decides whether to mark the image data in the current processing unit as a target bitstream segment to be output.

[0049] By employing the above-described vector product calculation method, this embodiment can efficiently and accurately determine the spatial relationship between the imaging area and the target area, ensuring the accuracy of data selection.

[0050] Example 4: Optionally, this embodiment adds a preprocessing step and a geometric correction step to the basis of embodiment 2 to generate a higher-level product.

[0051] The method in this embodiment adds the following steps before and after the steps in embodiment 2: Step A: Preprocessing (performed before the parallel streaming processing step) Before receiving large-scale data, based on the target area information in the mission plan, the system preloads the radiometric correction coefficients (such as absolute radiometric calibration coefficients Gain and Offset) corresponding to the target area from the radiometric correction coefficient database, as well as reference images (such as DOM orthophotos of the area) or digital elevation model (DEM) data for geometric fine correction; these data are loaded into a cache for quick retrieval in subsequent processing.

[0052] Step B: Radiometric correction (performed after the Grade 0 slice product is generated) After step 4 of Example 2 generates a Level 0 slice product, the system immediately starts the radiometric correction processing thread. This thread reads the image pixel value DN from the Level 0 slice product and, in conjunction with the radiometric correction coefficients pre-loaded in step A, performs radiometric correction calculations on each pixel to obtain a physically meaningful radiance value L or reflectance value ρ. The calculation formula is as follows: Gain and Offset are calibration coefficients for this sensor and band; this step converts the original DN value image into a radiometric image, completing the radiometric correction.

[0053] Step C: Rapid geometric correction (performed after radiometric correction) To generate a geocoded Level 1 product, geometric correction is required. Traditional methods require waiting for auxiliary parameters from the entire scene data, but this embodiment utilizes satellite orbit and attitude parameters already resolved during parallel streaming processing, covering the Level 0 tile range. Assuming the Level 0 tile covers data from time period t_a to t_b, the parallel streaming processing has already resolved satellite positions [X_s(t), Y_s(t), Z_s(t)]^T and attitude R(t) at multiple moments within that time period. For any row of pixels within the tile, its imaging time t is known. A method of predicting the global orientation using local parameters is employed to model the exterior orientation elements within the tile range. For example, a polynomial model can be used to fit the changes in satellite position over time. Using the precise satellite positions at several discrete moments as observations, the polynomial coefficients a_x0, a_x1, ... are calculated using the least squares method. Similarly, the attitude parameters can also be fitted with polynomials. Then, using these fitted continuous functions, combined with the imaging geometric model and the DEM data loaded in step A, a strict correspondence between each image point and the ground point is established. Through resampling (such as the nearest neighbor method or bilinear interpolation), the radiometrically corrected image is projected onto a map coordinate system (such as UTM projection), ultimately generating a Level 1 tile product with geographic coordinate information.

[0054] Through the above steps, this embodiment not only quickly generates a Level 0 slice, but also further completes radiometric and geometric corrections on this basis, directly outputting a Level 1 product that can be used directly by the user, thus shortening the processing chain from the raw bitstream to the application-level product.

[0055] Example 5 Optionally, this embodiment is a further refinement of the data decryption and decompression steps in Embodiment 2.

[0056] In this embodiment, in the parallel streaming processing of step 3, the decryption and decompression processes are performed in the following more detailed manner: Detailed steps of decryption: (1) Frame synchronization and invalid frame rejection: The processing unit first performs a frame synchronization header search on the received raw code stream data packets; each data frame starts with a fixed synchronization word (such as "1ACFFC1D"); if the synchronization word cannot be found for several consecutive frames, it is determined that the frame is out of sync, the error is recorded and resynchronization is attempted; for successfully synchronized frames, the error flag bit in the frame header is checked; if the flag bit indicates that an uncorrectable error occurred in the frame during transmission (such as CRC check failure), the frame is marked as an invalid frame and discarded directly.

[0057] (2) Decryption rule identification: The frame header contains an encryption method identification field; this field indicates what encryption algorithm was used for the data in this frame (e.g., Algorithm A: Chinese national standard SM4, Algorithm B: AES-256); the system selects the corresponding decryption algorithm from the decryption library based on this identifier.

[0058] (3) Key acquisition and decryption: Based on information such as satellite identifier, mission identifier and frame count, the corresponding decryption key is read from the hardware security module; then, the selected decryption algorithm and key are used to decrypt the effective data payload part after the frame header; after decryption, the encrypted header field in the frame header is removed to recover the plaintext compressed data stream.

[0059] (4) Output: Reassemble the decrypted data frame into a data packet and pass it to the next processing unit.

[0060] The detailed steps for decompression are as follows: (1) Compression type identification: The decrypted data packet contains an auxiliary data area and an image data area; a specific field in the auxiliary data area identifies the compression algorithm type used in the image data area (such as lossless compression JPEG-LS, or lossy compression CCSDS-IDC algorithm). (2) Decompression execution: Based on the identified compression type, the corresponding decompression algorithm library is called; If it is JPEG-LS compression, then it is decoded line by line according to the JPEG-LS standard to reconstruct the original pixel values; If CCSDS-IDC compression is used, the image data needs to be recovered by first performing entropy decoding, then performing inverse discrete wavelet transform, and finally performing bit plane reconstruction according to the CCSDS 122.0-B-1 standard. (3) Data separation: After decompression, the originally mixed data streams are clearly separated; auxiliary data (such as timecode, attitude data, satellite positioning data, etc.) are encapsulated into structured auxiliary data file formats (such as XML or binary header files); image data are organized into pixel matrices in the original band order (such as BGR or panchromatic bands); (4) Output: The separated auxiliary data and image data are passed as the output of the current data packet to the processing unit that selects the original code stream data segment.

[0061] Through the detailed description of the decryption and decompression steps above, this embodiment ensures that the method can adapt to various encryption and compression standards, and has good compatibility and robustness.

[0062] In summary, the present invention, through the above-described multiple embodiments, details the entire process from raw bitstream reception, target location prediction, parallel streaming processing to product generation; these embodiments demonstrate the effectiveness and advancement of the method of the present invention in improving the efficiency of ground processing of satellite remote sensing data and achieving rapid screening of target data.

[0063] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

Claims

1. A method for rapid screening of raw load bitstreams based on a track extrapolation model, characterized in that: Includes the following steps: Data stream reception steps: The ground system receives the raw bit stream of the downlink payload from the satellite in real time; Target bitstream location prediction step: Based on the pre-built imaging geometry model and the extrapolated imaging arc precise trajectory information, predict the starting position of the original bitstream data containing the target geographical region in the original bitstream of the payload; Parallel streaming processing steps: Using data packets as processing units, starting from the starting position, decryption processing, decompression processing, and original bitstream data segment selection processing are performed in parallel on multiple data packets to extract the target bitstream segment corresponding to the target geographical region; Product output steps: Based on the target bitstream segment, generate and output the level 0 slice product.

2. The method for fast screening of raw load bitstream based on orbit extrapolation model according to claim 1, characterized in that: The target bitstream location prediction step further includes: Construct the imaging geometric model that includes satellite orbital parameters, attitude parameters, and sensor imaging parameters; Based on the satellite's predetermined mission plan or real-time received auxiliary data, the precise orbital information covering the imaging time period of the target geographical area is extrapolated. By combining the imaging geometry model and the precise orbit information, the corresponding pixel position of the target geographic area on the satellite imaging focal plane is calculated, and then the frame number range or timestamp range of the original bitstream containing the pixel position data in the whole orbit data is predicted to determine the starting position.

3. The method for rapid screening of raw load bitstream based on orbit extrapolation model according to claim 1, characterized in that: In the parallel streaming processing step, the data packet consists of a fixed number of data frames, and the decryption processing, decompression processing, and raw bitstream data segment selection processing are executed in parallel on multiple processing units in a pipeline manner.

4. The method for fast screening of raw load streams based on an orbit extrapolation model according to claim 1, characterized in that: The raw bitstream data segment selection process in the parallel streaming processing step further includes: Analyze the auxiliary data in the current processing unit to obtain real-time positioning information; Based on the real-time positioning information, update the imaging geometric model or orbit extrapolation result, and determine whether the imaging area corresponding to the current processing unit intersects with the target geographical area; If an intersection exists, the image data portion in that processing unit is marked as part of the target bitstream segment.

5. The method for fast screening of raw load streams based on an orbit extrapolation model according to claim 4, characterized in that: The step of determining whether the imaging area corresponding to the current processing unit intersects with the target geographical area specifically includes: For point targets, the vector product method is used to determine whether the point target is located within the quadrilateral formed by the latitude and longitude coordinates of the four corner points of the imaging area corresponding to the current processing unit; For a target area, the vector product method is used to determine whether the boundary points of the target area fall within the quadrilateral formed by the latitude and longitude coordinates of the four corner points of the imaging area corresponding to the current processing unit.

6. The method for fast screening of raw load bitstream based on orbit extrapolation model according to claim 1, characterized in that: The decryption process includes: identifying and removing invalid frames, decrypting valid data according to preset decryption rules and a key library, and removing the encryption header.

7. The method for fast screening of raw load bitstream based on orbit extrapolation model according to claim 1, characterized in that: The decompression process includes: decompressing the decrypted data according to a preset decompression algorithm, and separating the decompressed data stream into auxiliary data and image data.

8. The method for fast screening of raw load streams based on an orbit extrapolation model according to claim 1, characterized in that: In the product output step, the level 0 slice product includes the target bitstream segment and its corresponding auxiliary data.

9. The method for fast screening of raw load bitstream based on orbit extrapolation model according to claim 1, characterized in that: The method also The process includes a preprocessing step: prior to the parallel streaming processing step, radiometric correction coefficients and / or reference images corresponding to the target geographic area are preloaded.

10. The method for fast screening of raw load bitstream based on orbit extrapolation model according to claim 1, characterized in that: The method further includes a geometric correction step: using some satellite parameters that are analyzed in real time in the parallel streaming processing step, the system performs geometric correction on the level 0 slice product by using a local parameter to predict the global value, and generates a level 1 slice product.

Images