Method and system for improving geometric positioning precision of satellite-borne SAR (Synthetic Aperture Radar) ocean image

Abstract

The invention provides a satellite-borne SAR ocean image geometric positioning precision improving method and system, and the method comprises the steps: collecting AIS data at the same imaging moment as an SAR satellite and in an imaging region range, correcting the orbit position of the satellite-borne SAR satellite according to the actual imaging moment of a ship, and fusing the AIS data with the SAR satellite data, thereby improving the geometric positioning precision of the satellite-borne SAR ocean image. The method comprises the following steps: optimizing a ship to obtain a matching point pair of an SAR image position and an AIS position of the ship, correcting an RD model error by taking the AIS position as a control point according to an optimized matching point pair result and an orbit position of a satellite-borne SAR satellite corresponding to an actual imaging moment of each ship, and performing iterative solution again to obtain new geographic coordinate information of each image pixel value, finally, a system geometric correction image product with improved positioning precision is obtained, and the problem that in the prior art, ship positioning precision is not high only depending on SAR images obtained by SAR satellites is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of spaceborne synthetic aperture radar (SAR) image processing, and in particular to a method and system for improving the geometric positioning accuracy of spaceborne SAR ocean images. Background Art

[0002] Spaceborne SAR (Synthetic Aperture Radar, SAR) is an active remote sensing device that uses radar imaging technology to monitor the sea surface, can obtain high-resolution images, and identify the position and outline of ships. AIS (Automatic Identification System, AIS) is an automatic identification signal automatically sent by ships, containing information such as location and speed, but it depends on whether the ship's equipment is turned on and the signal coverage range. The relationship between the two is as follows: (1) Ship inspection and identity verification SAR wide-area monitoring: Spaceborne SAR can scan the sea surface over a large area and detect ships that do not turn on AIS (such as illegal fishing vessels and smuggling ships). These ships may evade supervision by turning off AIS.

[0003] AIS identity matching: Through the ship identity and trajectory data provided by AIS, the targets detected by SAR can be quickly verified as legitimate ships, reducing misjudgments (such as distinguishing merchant ships from icebergs and sea clutter).

[0004] (2) Dark Target Identification Detecting "invisible" ships: Some ships deliberately turn off AIS or falsify information (such as illegal fishing and piracy). SAR can independently detect these targets through radar echoes, filling the regulatory loopholes of AIS.

[0005] Data fusion improves accuracy: By overlaying SAR images with AIS data, it is possible to verify whether the trajectory reported by the vessel is true (for example: AIS shows the vessel is at point A, but SAR detects it is at point B).

[0006] The accuracy of target position in remote sensing images is one of the basic elements of target information. High-precision positioning information can significantly improve the application efficiency of remote sensing images. The error sources that affect the positioning accuracy of satellite-borne SAR images include orbit error, slant range measurement error and elevation error. The orbit error mainly depends on the accuracy of the navigation and positioning equipment used on the satellite; the slant range measurement error is mainly affected by errors such as atmospheric propagation delay, sampling time error and channel delay; the elevation error is mainly the error of the target height relative to the surface of the earth model, which mainly depends on the accuracy of the digital elevation model (DEM) used in the processing system. Commercial satellite-borne SAR satellites are subject to factors such as satellite weight, equipment cost control and lack of geometric calibration. In the absence of control points, the positioning accuracy of commercial satellite-borne SAR images is usually only in the hundreds of meters, which seriously limits its application efficiency in ocean observation.

[0007] The ground processing system of spaceborne SAR usually performs imaging processing, radiation correction and system geometric correction to obtain standardized image products of various levels. Among them, the system geometric correction (used to eliminate the inherent geometric distortion caused by the SAR imaging principle (slant range projection, Doppler frequency shift); SAR is based on slant range projection (non-vertical observation), and the original image has geometric deformation in the range (Range) and azimuth (Azimuth) (such as perspective contraction and overlap), and the geometric correction converts it into an orthographic projection (map coordinate system)) usually uses the Range Doppler (RD) model to correct the image into a geocoded image to obtain a system geometric correction data product. The RD model consists of three nonlinear equations: the slant range equation, the Doppler center frequency equation, and the earth model. It uses satellite ephemeris parameters and SAR processing parameters (the deviation between the actual satellite orbit and the ideal orbit (such as altitude and speed changes), and the fluctuation of attitude angles (pitch, roll, yaw) will cause image distortion, which needs to be corrected using ephemeris and attitude data). The exact solution of the equation group is obtained by iterative solution or Ferrari solution, and the geographic coordinate position information of each pixel in the image is obtained to complete the production of system geometric correction products. The equation group of the RD model is as follows: Doppler equation: The Doppler center frequency between SAR and target can be obtained from the direct relative motion relationship between SAR and target. Since the Doppler center frequency used in imaging processing is consistent with the Doppler center frequency used in the geometric correction equation, the influence of Doppler center frequency on positioning accuracy can be ignored.

[0008] (A) Slant range equation: SAR has precise ranging capability. According to the sampling parameters, the slant range value of each sample can be accurately obtained. When the spaceborne SAR performs imaging tasks, errors such as atmospheric propagation delay, sampling time error and channel delay are ultimately reflected in the slant range error.

[0009] (B) Earth model: The earth can usually be equivalent to an ellipsoid, and the target is a point on this ellipsoid. Based on this condition, the constraint equation of the target's three-dimensional position can be obtained.

[0010] (C) in, represents the Doppler center frequency, represents the microwave wavelength of spaceborne SAR, is the distance between the SAR and the ground target, and Represents the position and velocity vector of the SAR satellite using the GPS's own WGS84 earth-fixed coordinate system; and The position and velocity vectors of ground targets are usually represented using the WGS84 earth-fixed coordinate system. and Represents the major and minor axes of the ideal Earth ellipsoid model.

[0011] In order to improve the positioning accuracy of spaceborne SAR images, two methods, geometric calibration and orthorectification, are usually used. Geometric calibration is that the SAR satellite observes and images the ground calibration field multiple times, and uses the high-precision control point (GCP) data corresponding to the accurately measured ground corner reflector to obtain the geometric calibration coefficient, correct the systematic error of the spaceborne SAR positioning model, and thus improve the uncontrolled positioning accuracy of the geometric correction product of the spaceborne SAR system. Spaceborne SAR orthorectification uses geometric models, DEM and GCP data to correct various geometric errors and distortions in the imaging process, obtain more advanced orthorectified image products, and further improve the geometric quality of spaceborne SAR images. During the on-orbit operation of spaceborne SAR satellites, each time they perform imaging tasks, they are inevitably affected by factors such as the space environment, equipment temperature, platform stability and orbit control, resulting in the ground calibration coefficient being unable to fully compensate for the positioning error. In addition, the atmospheric delay error correction depends on atmospheric parameter products (European Center for Medium-Range Numerical Weather Forecasts ECMWF and National Center for Environmental Prediction NCEP of the United States), which have time delays and uncertainties, and also have large time delays themselves. However, the measured atmospheric parameters require a lot of manpower and material resources, which is almost impossible for most commercial satellite companies to implement. Although the orthorectification method can correct geometric errors, it relies on high-precision GCP points and has strict requirements on the number and distribution of GCP points. In ocean scenes, GCP points are difficult to obtain, which makes it difficult for the orthorectification method to meet the real-time processing requirements of the spaceborne SAR ground system. Summary of the invention

[0012] The purpose of the present invention is to overcome the above technical deficiencies and provide a method and system for improving the geometric positioning accuracy of satellite-borne SAR ocean images, so as to solve the problem in the related art that the ship positioning accuracy is low only by relying on SAR images obtained by SAR satellites.

[0013] In order to achieve the above technical objectives, the present invention adopts the following technical solutions: According to a first aspect of the present invention, a method for improving the geometric positioning accuracy of spaceborne SAR ocean images is provided, comprising: Step S11, formulate the future imaging mission plan and data transmission reception mission of the spaceborne SAR satellite system, determine the satellite imaging area range and the specific imaging time according to the imaging mission plan, and after the SAR satellite completes imaging according to the imaging mission plan, collect AIS data within the corresponding time and area through the AIS data service provider; Step S12, according to the data transmission receiving task, pre-processing the original code stream data received by the data transmission station of the spaceborne SAR satellite ground processing system to obtain a spaceborne SAR ocean image with different scenes; Step S13, performing ship detection on the satellite-borne SAR ocean image, and calculating the actual imaging time of each detected ship, so as to synchronize the time when fusing with the AIS data; Step S14, calculating the orbital position of the spaceborne SAR satellite corresponding to the actual imaging time of each ship based on the GPS data and the actual imaging time of each ship in combination with the dynamic model; Step S15: pre-matching the ships in the satellite-borne SAR ocean image and the AIS data according to the actual imaging time of each ship, to obtain a combination of the SAR image position and the AIS position of the ship; Step S16: Based on the global minimum mean square error, matching optimization is performed on the combination of the SAR image position and the AIS position to obtain an optimized matching point pair; Step S17: Based on the optimized matching point pairs and the orbital position of the spaceborne SAR satellite corresponding to the actual imaging time of each ship, the AIS position is used as a control point to correct the RD model error, and the new geographic coordinate information of each SAR image pixel value is re-solved to obtain a geometrically corrected image product with improved positioning accuracy.

[0014] According to a second aspect of the present invention, a system for improving geometric positioning accuracy of spaceborne SAR ocean images is provided, comprising: The data acquisition and processing module is used to formulate the future imaging mission planning and data transmission reception mission of the spaceborne SAR satellite system, determine the satellite imaging area range and the specific imaging time according to the imaging mission planning, and collect AIS data within the corresponding time and area range through the AIS data service provider after the SAR satellite completes imaging according to the imaging mission planning; It is also used to pre-process the original code stream data received by the data transmission station of the spaceborne SAR satellite ground processing system according to the data transmission reception task to obtain the spaceborne SAR ocean image with different scenes; A calculation module, used for detecting ships on the satellite-borne SAR ocean image and calculating the actual imaging time of each detected ship, so as to synchronize the time when fusing with the AIS data; It is also used to calculate the orbital position of the spaceborne SAR satellite corresponding to the actual imaging time of each ship based on the GPS data and the actual imaging time of each ship in combination with the dynamic model; A matching module is used to pre-match the ship in the satellite-borne SAR ocean image and the AIS data according to the actual imaging time of each ship, so as to obtain a combination of the SAR image position and the AIS position of the ship; An optimization module, used for performing matching optimization on the combination of the SAR image position and the AIS position based on a global minimum mean square error to obtain an optimized matching point pair; The correction module is used to correct the RD model error based on the optimized matching point pairs and the orbital position of the space-borne SAR satellite corresponding to the actual imaging time of each ship, using the AIS position as the control point, and re-solve to obtain the new geographic coordinate information of each SAR image pixel value to obtain a geometrically corrected image product with improved positioning accuracy.

[0015] The technical solution provided by the embodiments of the present invention may have the following beneficial effects: By collecting AIS data at the same imaging time and imaging area as the SAR satellite, and correcting the orbital position of the spaceborne SAR satellite according to the actual imaging time of the ship, the AIS data and the SAR satellite data are fused to obtain the matching point pairs of the SAR image position and the AIS position of the ship. According to the optimized matching point pair results and the orbital position of the spaceborne SAR satellite corresponding to the actual imaging time of each ship, the AIS position is used as the control point to correct the RD model error, and the iterative solution is re-performed to obtain the new geographic coordinate information of each image pixel value, and finally a system geometric correction image product with improved positioning accuracy is obtained, which solves the problem of low accuracy of ship positioning in the prior art that only relies on SAR images obtained by SAR satellites. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1is a flow chart of a method for improving the geometric positioning accuracy of spaceborne SAR ocean images according to an exemplary embodiment; Figure 2 It is a schematic block diagram of a system for improving the geometric positioning accuracy of space-borne SAR ocean images according to an exemplary embodiment. DETAILED DESCRIPTION

[0017] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0018] Embodiment 1 Figure 1 is a flow chart of a method for improving the geometric positioning accuracy of spaceborne SAR ocean images according to an exemplary embodiment. Figure 1 , the method comprising: Step S11, formulate the future imaging mission plan and data transmission reception mission of the spaceborne SAR satellite system, determine the satellite imaging area range and the specific imaging time according to the imaging mission plan, and after the SAR satellite completes imaging according to the imaging mission plan, collect AIS data within the corresponding time and area through the AIS data service provider; Step S12, according to the data transmission receiving task, pre-processing the original code stream data received by the data transmission station of the spaceborne SAR satellite ground processing system to obtain a spaceborne SAR ocean image with different scenes; Step S13, performing ship detection on the satellite-borne SAR ocean image, and calculating the actual imaging time of each detected ship, so as to synchronize the time when fusing with the AIS data; Step S14, calculating the orbital position of the spaceborne SAR satellite corresponding to the actual imaging time of each ship based on the GPS data and the actual imaging time of each ship in combination with the dynamic model; Step S15: pre-matching the ships in the satellite-borne SAR ocean image and the AIS data according to the actual imaging time of each ship, to obtain a combination of the SAR image position and the AIS position of the ship; Step S16: Based on the global minimum mean square error, the combination of the SAR image position and the AIS position is matched and optimized to obtain an optimized matching point pair; Step S17: Based on the optimized matching point pairs and the orbital position of the spaceborne SAR satellite corresponding to the actual imaging time of each ship, the AIS position is used as a control point to correct the RD model error, and the new geographic coordinate information of each SAR image pixel value is re-solved to obtain a geometrically corrected image product with improved positioning accuracy.

[0019] It should be noted that the technical solution provided in this embodiment is loaded and run in an electronic device in specific practice, and the electronic device includes: at least one processor; and a memory connected to the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method described in the embodiment. The electronic device is generally set in a SAR satellite ground receiving center.

[0020] It can be understood that the technical solution provided in this embodiment collects AIS data within the same imaging time and imaging area as the SAR satellite, and corrects the orbital position of the spaceborne SAR satellite according to the actual imaging time of the ship, and fuses the AIS data and the SAR satellite data to obtain the matching point pair of the SAR image position and the AIS position of the ship. According to the optimized matching point pair results and the orbital position of the spaceborne SAR satellite corresponding to the actual imaging time of each ship, the AIS position is used as a control point to correct the RD model error, and the iterative solution is re-performed to obtain new geographic coordinate information of each image pixel value, and finally a system geometric correction image product with improved positioning accuracy is obtained, which solves the problem of low accuracy of ship positioning in the prior art that only relies on SAR images obtained by SAR satellites.

[0021] It should be noted that in step S11, according to the orbit of the SAR satellite and the planned imaging ocean area position, the future imaging mission planning and data transmission reception mission of the spaceborne SAR satellite system are formulated. The AIS data service provider collects AIS data within the time and location range, and the AIS data time usually covers half an hour before and after the spaceborne SAR imaging time.

[0022] In order to facilitate understanding of the specific implementation of preprocessing the original code stream data received by the data transmission station of the spaceborne SAR satellite ground processing system according to the data transmission reception task in step S12 to obtain the spaceborne SAR ocean image with different scenes, the relevant technical terms are explained as follows: The coverage area obtained by a satellite scan is called a "scene", which is similar to a photo taken by a camera. The size of a scene depends on the satellite imaging mode, usually covering tens to hundreds of kilometers (for example, a Sentinel-1 scene is about 250 kilometers × 250 kilometers). When SAR satellite imaging, it is necessary to divide it according to satellite orbit parameters (such as latitude / longitude) or mission requirements to ensure seamless splicing between adjacent scenes.

[0023] Spaceborne SAR data needs to go through multiple levels of processing from raw signals to usable images, forming standardized products of different levels: Level 0 (raw data): Unprocessed radar echo signals (binary data packets) that cannot be used directly by users.

[0024] Level 1 (primary product): Single look complex image (SLC): contains phase and amplitude information for specialized analysis (such as interferometry).

[0025] Ground Range Image (GRD): After geometric correction, it can be directly used for target detection or mapping.

[0026] Level 2+ (Advanced Product): Further processed into thematic data, such as ocean wind fields, surface deformation maps, etc.

[0027] The "single-view product" downloaded by the user is an image that can be used directly (such as Level 1 GRD) without having to process it from the original signal. Each single-view product contains: Image data: A grayscale image reflecting the backscattering intensity of the surface.

[0028] Metadata: shooting time, satellite parameters (incident angle, polarization mode), geographical range, etc.

[0029] Specifically, image products of all levels are obtained through data analysis, data segmentation and standard scene product production. Among them, the system geometry correction product adopts the RD model, and the geographic coordinate information of each image pixel value is obtained through iterative solution. The satellite orbit data of at least 4 seconds during the imaging period is obtained based on the auxiliary data, and the data is usually in seconds.

[0030] It should be noted that, in step S13, a constant false-alarm rate (CFAR) or a deep learning method is used to perform ship detection on the satellite-borne SAR ocean image. Since the ship detection method is a prior art, it will not be described in detail in this embodiment.

[0031] In specific practice, the actual imaging time of each detected ship is calculated in step S13, including: According to the ship detection results, the row and column coordinates of all ships on the spaceborne SAR image are obtained from the spaceborne SAR ocean image. ,That and Corresponding to the pixel positions in azimuth and distance respectively, Indicates the number of detected ships; Get the latitude and longitude position values ​​of all ships from the spaceborne SAR ocean image ,in, Indicates longitude, Indicates latitude, Indicates elevation; Get the central imaging time of the scene image product from the metadata of the current foreground image product , the pulse repetition frequency prf of the echo signal, the azimuth starting time of the imaging process ; Calculate the actual imaging moment for each ship , .

[0032] It is understandable that the SAR imaging process involves signal delay and Doppler shift. Since the satellite is moving, the echoes received at different time points correspond to different position information. For stationary targets, their imaging positions can be determined by the geometric relationship between the satellite position and the beam. However, for moving ships, their movement may affect the echo delay and Doppler parameters, resulting in position offset or blur in the image.

[0033] Therefore, calculating the actual imaging time of each ship is to more accurately locate their position in the image. For example, if the ship is moving, its speed will cause the position in the SAR image to deviate from the actual position. At this time, knowing the specific imaging time point combined with the satellite's orbital parameters can correct this deviation or infer the speed and heading of the ship.

[0034] In addition, the actual position of each pixel in the SAR image is calculated by the propagation time of the radar signal and the position of the satellite. For stationary targets, this process is relatively straightforward, but moving targets introduce additional time-varying factors. Therefore, determining the actual imaging moment of the ship helps to accurately align its position, especially when it needs to be fused with other data sources (such as AIS data). Time synchronization is very important.

[0035] In specific practice, the orbital position of the spaceborne SAR satellite corresponding to the actual imaging time of each ship is calculated in step S14 according to the GPS data and the actual imaging time of each ship in combination with the dynamic model, including: According to the GPS data in the spaceborne SAR auxiliary data, the third-order polynomial method is used to perform parameter fitting, including: (1) in, , Indicates the orbital position of the SAR satellite at time t0 in the GPS data; Indicates the orbital position of the SAR satellite at time t1 in the GPS data; Indicates the orbital position of the SAR satellite at time t2 in the GPS data; Indicates the orbital position of the SAR satellite at time t3 in the GPS data; the first column in the matrix T is the number 1; According to the least squares method, the specific values ​​of each parameter in matrix A are calculated: (2) According to the actual imaging time of each ship , the orbital position of the spaceborne SAR satellite at the corresponding time is obtained by third-order fitting calculation : (3).

[0036] It is understandable that in the spaceborne SAR (synthetic aperture radar) system, using GPS data for orbit data fitting is the core technical means to ensure high-precision acquisition of satellite position, velocity and attitude parameters. This process directly affects the geometric positioning accuracy of SAR images, interferometric measurement capabilities and the reliability of mobile target detection.

[0037] In spaceborne SAR (synthetic aperture radar) systems, auxiliary data usually contains various parameters required for satellite operation, of which GPS data is an important component. The GPS receiver records the satellite's three-dimensional position (longitude, latitude, altitude) and speed in real time for subsequent imaging geometry correction and geolocation.

[0038] In the spaceborne SAR (synthetic aperture radar) system, using GPS data to fit orbital data is the core technical means to ensure high-precision acquisition of satellite position, velocity and attitude parameters. This process directly affects the geometric positioning accuracy, interferometric measurement capability and reliability of mobile target detection of SAR images. SAR synthesizes a virtual long-aperture antenna through the motion of the satellite platform, and its imaging and data processing are highly dependent on the temporal and spatial parameters of the satellite orbit.

[0039] Geometric positioning: The geographic coordinates of each pixel in the SAR image are determined by the radar wave propagation time and the satellite position. If the orbit error exceeds 1 meter, the image positioning error may exceed 10 meters (especially in low incidence angle scenarios).

[0040] GPS provides a real-time, high-frequency, and high-precision orbit measurement method for spaceborne SAR, and its advantages are irreplaceable: three-dimensional position (X / Y / Z) and velocity (Vx / Vy / Vz) directly cover all the degrees of freedom required for orbit fitting. The data output frequency of a typical GPS receiver is 1 Hz~10 Hz, which is much higher than that of a ground measurement and control system (minute level), and can capture the instantaneous motion details of the satellite.

[0041] In specific practice, in step S15, according to the actual imaging time of each ship, the ship in the satellite-borne SAR ocean image and the AIS data are pre-matched to obtain a combination of the SAR image position and the AIS position of the ship, including: The AIS data is classified according to the ship identification code and timestamp, and the data with heading changes in the ship's AIS dynamic information is cleared, and the data whose AIS data time cannot cover the spaceborne SAR imaging time is cleared; According to the timestamp and position of the ship AIS data, the second-order fitting method is used to calculate the imaging time of the ship in the spaceborne SAR image. AIS latitude and longitude position values ,in, Indicates longitude, Indicates latitude, Indicates the number of vessels that meet the AIS criteria; The latitude and longitude position value P of the AIS i And the latitude and longitude position values ​​R of all ships detected on the SAR image i Perform spatial correlation to obtain the combination of the SAR image position and AIS position of the ship ,in Indicates the number of ships in the SAR image that can be matched with the AIS signal according to the spatial association rules, It represents the number of AIS signals that can be matched and associated with ship j according to the spatial association rule. L is a positive integer greater than 2. If a ship in the SAR image fails to match the AIS data, the detection result of the ship is deleted.

[0042] It should be noted that according to the spatial rule with a tolerance value of 1.5 times, each ship may be associated with multiple AIS signals, and L here represents the number of AIS signals associated with ship j.

[0043] It can be understood that the inventive concept of the present invention is: when processing satellite SAR image data, correlating AIS data to verify or analyze the activities of the ship, whether the AIS data position at a specific time point (the actual imaging moment of the ship) matches the ship detection result in the SAR image.

[0044] In this embodiment, the latitude and longitude coordinates of the ship are detected based on the satellite-borne SAR image. ,The ships in the image and the corresponding AIS longitude and latitude position value data are spatially associated by setting a spatial region with a tolerance range, and the tolerance value is set to 1.5 times of the positioning accuracy index of the space-borne SAR image.

[0045] According to the AIS information standard, the reporting interval for static information such as the information identification code and MMSI number is 6 minutes, while the reporting interval for dynamic information such as ground speed, ship position longitude and latitude, and heading ranges from 2 seconds to 3 minutes.

[0046] Cleaning AIS data is to improve the accuracy of subsequent analysis so as to conduct time-space matching analysis. Clean AIS data can help eliminate false alarms, such as position deviations caused by heading changes, or data unavailability caused by time mismatches, and ensure that the time windows of AIS data and SAR images are completely consistent, so as to accurately associate the data of the two and improve the reliability of detection or verification. For example, if the SAR image is taken at a specific moment, and the AIS data is not recorded before and after that moment, then these data need to be eliminated to avoid introducing noise.

[0047] In practice, the judgment of heading changes requires setting a threshold, for example, a heading change exceeding a certain angle is considered a valid change to avoid misjudgment due to data noise. At the same time, time coverage requires defining a time interval, such as a few minutes before and after the SAR imaging moment, to accommodate the time deviation of AIS data, because AIS data may have time delays or update frequency issues.

[0048] In specific practice, in step S16, matching optimization is performed on the combination of the SAR image position and the AIS position based on the global minimum mean square error to obtain an optimized matching point pair, including: Calculates the combination of the SAR image position and the AIS position of the vessel The distance between the SAR image position and the AIS position of each ship in the longitude direction ; According to the distance The average value of the SAR image position and the AIS position is matched and optimized by using the global minimum mean square error association algorithm.

[0049] In the spaceborne SAR imaging processing, when the ship target moves in the direction perpendicular to the satellite orbit, the imaging position of the ship in the spaceborne SAR image will be offset along the satellite orbit direction. The offset is related to the ship's movement speed, movement direction and radar incident angle.

[0050] Preferably, the calculation of the combination of the SAR image position and the AIS position of the vessel The distance between the SAR image position and the AIS position of each ship in the longitude direction , specifically: (4) S jk = COS(R j -Lon k )(5) in, R represents the distance in longitude between the jth ship in the SAR image that can be associated with the AIS and the kth ship in the AIS position that can be associated with the SAR image; j Indicates the longitude value of the jth ship that can be associated with AIS in the SAR image, Lon k Indicates the longitude value of the kth ship in the AIS position that can be associated with the SAR image.

[0051] Preferably, the distance The average value of , using the global minimum mean square error association algorithm, is used to match and optimize the combination of the SAR image position and the AIS position, including: The global minimum mean square error association algorithm is used to calculate the minimum mean square value of the distance deviation of all ships in the longitude direction. The designed objective function is as follows: (6) in, Indicates that all the The average value of is a binary variable, 0 means no association, 1 means association, in a set of rough matching There is at most one associated value in Exceeds the number of matching pairs in this group When The value is 0; Indicates that all ships are roughly matched The maximum value of The Munkres algorithm is used to perform global minimum mean square error point matching calculation to obtain new matching point pairs between the SAR image position and the AIS position. ,in , represents the number of ships that achieved matching; calculate The Euclidean distance between the SAR image position and the AIS position of the vessel : (7) Calculate all matching point pairs The mean and standard deviation , judge whether the associated matching point pair exceeds the threshold. If it exceeds the threshold, it means that the associated matching is invalid and the matching point is deleted. The threshold range is , and obtain the matching point pairs after spatiotemporal correlation optimization .

[0052] It can be understood that the optimization method based on the global minimum mean square error (MMSE) applied to the matching of SAR images and AIS (Automatic Identification System) positions can significantly improve the accuracy and reliability of multi-source data fusion.

[0053] MMSE finds the optimal matching combination between the ship position detected in the SAR image and the dynamic position reported by AIS by minimizing the mean square value of the global position error. This method is statistically unbiased and effective, and can eliminate the impact of random noise (such as SAR imaging error, AIS transmission delay) on the matching results. For example, if the detection coordinates of a ship in the SAR image are A, and AIS reports multiple candidate positions B1, B2, B3, ...........Bn in the same period, MMSE calculates the global mean square error and selects the AIS record with the smallest error as the matching object.

[0054] In addition, in dense sea areas, SAR may detect multiple ships at the same time (such as ports or waterways), while AIS data may contain hundreds of ship dynamic information. MMSE can avoid local optimal traps (such as matching only the nearest neighbor AIS target) through global optimization and achieve a global optimal solution for many-to-many matching. For example, if SAR images detect 5 ships and AIS reports 10 ships in the same area, MMSE can avoid mismatching a SAR ship to multiple AIS targets or missing low-speed ships by jointly optimizing all possible matching combinations.

[0055] In specific practice, in step S17, based on the optimized matching point pair and the orbital position of the spaceborne SAR satellite corresponding to the actual imaging time of each ship, the AIS position is used as a control point to correct the RD model error, and the new geographic coordinate information of each SAR image pixel value is obtained again to obtain a geometrically corrected image product with improved positioning accuracy, including: The positioning error of spaceborne SAR images can be divided into range positioning error and azimuth error according to the satellite imaging direction. According to the RD model, the errors in these two dimensions can be expressed by the initial slant range error in the range direction and the error in the starting time in the azimuth direction.

[0056] The slope distance equation of each pixel in the RD model is expressed as the following formula (8):

[0057] in, Indicates the initial slant distance used in slant distance calculation and the near distance value in imaging processing, which is obtained from the metadata of the current foreground image product; is the speed of light constant, is the number of sampling points in the distance direction; Indicates the sampling frequency of the echo signal, which is obtained from the metadata of the current foreground image product; Represents the range error, including the delay error introduced by atmospheric propagation, the system sampling time error and the channel delay error. This error value takes a fixed value in a single imaging task; The azimuth starting time error formula is as follows: (9) in, Indicates the starting time of the orientation in the imaging process, which is obtained from the metadata of the current foreground image product; is the azimuth pulse count; Indicates the time error in azimuth; According to formula (3), the orbital position error of the spaceborne SAR satellite is Expressed as a first-order function of the azimuth time error: (10) Combining formula (8) and formula (10), we can get the initial slant range error in the range direction: and the azimuth starting time error Equation (11):

[0058] According to equation (11), calculate and The RD model is modified according to the value of

[0059] In specific practice, according to equation (11), and The RD model is modified based on the value of Step S171, obtaining calculation parameters, the calculation parameters include: the longitude and latitude coordinates of the AIS are expressed as the coordinate position of the ship in the SAR image , corresponding to the column coordinates of the ship on the spaceborne SAR image , and the parameters in the metadata and , and the actual imaging time of the vessel The corresponding orbital position of the spaceborne SAR satellite ; Step S172: Use the matching point pairs optimized by time-space correlation ,calculate and The RD model is modified by express The number of matches in , including: when , using formula (8), substitute , , , and calculate , and take the average value as the range slant range error , default azimuth time error is 0; when , using formula (10), we can obtain equations, set the initial and are all 0, and the Newton iteration method is used to calculate using the first-order derivative and ; Determine whether the error is less than the preset threshold through iterative calculation ,when When , the iteration is terminated, otherwise, the process goes to step S171 to continue the iterative calculation; Step S173: and As a result, the RD model equations are updated, and the corrected system geometry correction algorithm equations include: (12) (13) (14) The geographic coordinate position information of each pixel in the image is obtained again through iterative solution or the Ferrari solution, and finally a system geometric correction image product with improved positioning accuracy is obtained.

[0060] It can be seen from the description of the above technical solutions that the technical solution provided by this embodiment improves the quality of ocean images in satellite-borne SAR ground processing and solves the problem of insufficient calibration parameters and control points of commercial satellite-borne SAR satellites. According to the dynamic and static information in AIS and the ship detection information of satellite-borne SAR images, a high-precision association between AIS signals and ship SAR images is achieved through a multi-strategy spatiotemporal correlation matching method. At the same time, the RD correction model in the satellite-borne SAR ground processing system is corrected using the AIS position information to obtain the two-dimensional error parameters in azimuth and range, and the RD model equation is reconstructed to achieve the timeliness and reliability of the accuracy improvement of the geometric correction products of the ground processing system.

[0061] First, the spatiotemporal correlation matching process of spaceborne SAR image ships and AIS data was improved. The position of the ship image is offset due to the motion of the ship target, which is the cause of the spaceborne SAR imaging mechanism. The various existing methods basically use motion parameter calculation to compensate and correct the ship position. By designing a preliminary matching strategy based on the distance in the longitude direction and combining it with the global minimum mean square error optimization algorithm, the matching efficiency and accuracy of AIS data and SAR image ships are significantly improved, solving the matching problem in multi-target scenarios.

[0062] Secondly, the modified RD model algorithm for geometric correction of the spaceborne SAR system targets the random and systematic deviations that cannot be eliminated in the spaceborne SAR imaging process. It uses the meter-level precision position information of AIS to calibrate the initial slant range error in the distance and the starting time error in the azimuth in the RD model, reconstructs the RD model equation, and effectively improves the geometric correction accuracy of the spaceborne SAR ground processing system.

[0063] Compared with the prior art, the technical solution provided by this embodiment includes at least the following beneficial effects: (1) By introducing AIS data as control points, the geometric positioning accuracy of spaceborne SAR ocean images is significantly improved, solving the problems of no control points and insufficient calibration parameter updates.

[0064] (2) Based on the existing business process of the spaceborne SAR ground processing system, the geometric correction algorithm of the spaceborne SAR ground processing system was optimized by importing AIS data and optimizing the geometric correction algorithm without adding hardware facilities, thereby improving the processing efficiency and reliability of ocean images.

[0065] (3) This method provides technical support for the application of high-resolution spaceborne SAR data in the fields of ocean monitoring and target identification, and has broad application prospects.

[0066] Embodiment 2 Figure 2is a schematic block diagram of a system 100 for improving the geometric positioning accuracy of spaceborne SAR ocean images according to an exemplary embodiment. Figure 2 , the system 100 comprises: The data acquisition and processing module 101 is used to formulate the future imaging mission planning and data transmission reception mission of the spaceborne SAR satellite system, determine the satellite imaging area range and the specific imaging time according to the imaging mission planning, and collect AIS data within the corresponding time and area range through the AIS data service provider after the SAR satellite completes imaging according to the imaging mission planning; It is also used to pre-process the original code stream data received by the data transmission station of the spaceborne SAR satellite ground processing system according to the data transmission reception task to obtain the spaceborne SAR ocean image with different scenes; A calculation module 102 is used to detect ships on the satellite-borne SAR ocean image and calculate the actual imaging time of each detected ship for time synchronization when fusing with the AIS data; It is also used to calculate the orbital position of the spaceborne SAR satellite corresponding to the actual imaging time of each ship based on the GPS data and the actual imaging time of each ship in combination with the dynamic model; A matching module 103 is used to pre-match the ship in the satellite-borne SAR ocean image and the AIS data according to the actual imaging time of each ship, so as to obtain a combination of the SAR image position and the AIS position of the ship; An optimization module 104 is used to perform matching optimization on the combination of the SAR image position and the AIS position based on a global minimum mean square error to obtain an optimized matching point pair; The correction module 105 is used to correct the RD model error based on the optimized matching point pair and the orbital position of the space-borne SAR satellite corresponding to the actual imaging time of each ship, using the AIS position as a control point, and re-solve to obtain new geographic coordinate information of each SAR image pixel value to obtain a geometrically corrected image product with improved positioning accuracy.

[0067] It should be noted that the technical solution provided in this embodiment is loaded and run in an electronic device in specific practice, and the electronic device includes: at least one processor; and a memory connected to the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method described in the embodiment. The electronic device is generally set in a SAR satellite ground receiving center.

[0068] The implementation methods of the above modules refer to the implementation methods of the corresponding steps in Example 1, which will not be repeated in this example.

[0069] It can be understood that the technical solution provided in this embodiment collects AIS data within the same imaging time and imaging area as the SAR satellite, and corrects the orbital position of the spaceborne SAR satellite according to the actual imaging time of the ship, and fuses the AIS data and the SAR satellite data to obtain the matching point pair of the SAR image position and the AIS position of the ship. According to the optimized matching point pair results and the orbital position of the spaceborne SAR satellite corresponding to the actual imaging time of each ship, the AIS position is used as a control point to correct the RD model error, and the iterative solution is re-performed to obtain new geographic coordinate information of each image pixel value, and finally a system geometric correction image product with improved positioning accuracy is obtained, which solves the problem of low accuracy of ship positioning in the prior art that only relies on SAR images obtained by SAR satellites.

[0070] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0071] If the integrated units in the above embodiments are implemented in the form of software functional units and sold or used as independent products, they can be stored in the above computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling one or more computer devices (which may be personal computers, servers or network devices, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application.

[0072] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0073] In the several embodiments provided in the present application, it should be understood that the disclosed client can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0074] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0075] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0076] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for improving the geometric positioning accuracy of spaceborne SAR ocean images, characterized in that: include: Step S11, formulate the future imaging mission plan and data transmission reception mission of the spaceborne SAR satellite system, determine the satellite imaging area range and the specific imaging time according to the imaging mission plan, and after the SAR satellite completes imaging according to the imaging mission plan, collect AIS data within the corresponding time and area through the AIS data service provider; Step S12, according to the data transmission receiving task, pre-processing the original code stream data received by the data transmission station of the spaceborne SAR satellite ground processing system to obtain a spaceborne SAR ocean image with different scenes; Step S13, performing ship detection on the satellite-borne SAR ocean image, and calculating the actual imaging time of each detected ship, so as to synchronize the time when fusing with the AIS data; Step S14, calculating the orbital position of the spaceborne SAR satellite corresponding to the actual imaging time of each ship based on the GPS data and the actual imaging time of each ship in combination with the dynamic model; Step S15: pre-matching the ships in the satellite-borne SAR ocean image and the AIS data according to the actual imaging time of each ship, to obtain a combination of the SAR image position and the AIS position of the ship; Step S16: Based on the global minimum mean square error, matching optimization is performed on the combination of the SAR image position and the AIS position to obtain an optimized matching point pair; Step S17: Based on the optimized matching point pairs and the orbital position of the spaceborne SAR satellite corresponding to the actual imaging time of each ship, the AIS position is used as a control point to correct the RD model error, and the new geographic coordinate information of each SAR image pixel value is re-solved to obtain a geometrically corrected image product with improved positioning accuracy.

2. The method according to claim 1, characterized in that The actual imaging time of each detected ship is calculated in step S13, including: According to the ship detection results, the row and column coordinates of all ships on the spaceborne SAR image are obtained from the spaceborne SAR ocean image. ,That and Corresponding to the pixel positions in azimuth and distance respectively, Indicates the number of detected ships; Get the latitude and longitude position values ​​of all ships from the spaceborne SAR ocean image ,in, Indicates longitude, Indicates latitude, Indicates elevation; Get the central imaging time of the scene image product from the metadata of the current foreground image product , the pulse repetition frequency prf of the echo signal, the azimuth starting time of the imaging process ; Calculate the actual imaging moment for each vessel , .

3. The method according to claim 2, characterized in that In step S14, the orbital position of the spaceborne SAR satellite corresponding to the actual imaging time of each ship is calculated based on the GPS data and the actual imaging time of each ship in combination with the dynamic model, including: According to the GPS data in the spaceborne SAR auxiliary data, the third-order polynomial method is used to perform parameter fitting, including: （1） in, , Indicates the orbital position of the SAR satellite at time t0 in the GPS data; Indicates the orbital position of the SAR satellite at time t1 in the GPS data; Indicates the orbital position of the SAR satellite at time t2 in the GPS data; Indicates the orbital position of the SAR satellite at time t3 in the GPS data; the first column in the matrix T is the number 1; According to the least squares method, the specific values ​​of each parameter in matrix A are calculated: （2） According to the actual imaging time of each ship , the orbital position of the spaceborne SAR satellite at the corresponding time is obtained by third-order fitting calculation : （3）。 4. The method according to claim 3, characterized in that In step S15, according to the actual imaging time of each ship, the ship in the satellite-borne SAR ocean image and the AIS data are pre-matched to obtain a combination of the SAR image position and the AIS position of the ship, including: The AIS data is classified according to the ship identification code and timestamp, and the data with heading changes in the ship's AIS dynamic information is cleared, and the data whose AIS data time cannot cover the spaceborne SAR imaging time is cleared; According to the timestamp and position of the ship AIS data, the second-order fitting method is used to calculate the imaging time of the ship in the spaceborne SAR image. AIS latitude and longitude position values ,in, Indicates longitude, Indicates latitude, Indicates the number of vessels that meet the AIS criteria; The latitude and longitude position value P of the AIS i And the latitude and longitude position values ​​R of all ships detected on the SAR image i Perform spatial correlation to obtain the combination of the SAR image position and AIS position of the ship ,in Indicates the number of ships in the SAR image that can be matched with the AIS signal according to the spatial association rules, It represents the number of AIS signals that can be matched and associated with ship j according to the spatial association rule. L is a positive integer greater than 2. If a ship in the SAR image fails to match the AIS data, the detection result of the ship is deleted.

5. The method according to claim 4, characterized in that In step S16, based on the global minimum mean square error, the combination of the SAR image position and the AIS position is matched and optimized to obtain an optimized matching point pair, including: Calculates the combination of the SAR image position and the AIS position of the vessel The distance between the SAR image position and the AIS position of each ship in the longitude direction ; According to the distance The average value of the SAR image position and the AIS position is matched and optimized by using the global minimum mean square error association algorithm.

6. The method according to claim 5, characterized in that The calculation is a combination of the SAR image position and the AIS position of the vessel The distance between the SAR image position and the AIS position of each ship in the longitude direction , specifically: （4） S jk = COS(R j -Lon k )(5) in, R represents the distance in longitude between the jth ship in the SAR image that can be associated with the AIS and the kth ship in the AIS position that can be associated with the SAR image; j Indicates the longitude value of the jth ship that can be associated with AIS in the SAR image, Lon k Indicates the longitude value of the kth ship in the AIS position that can be associated with the SAR image.

7. The method according to claim 6, characterized in that According to the distance The average value of , using the global minimum mean square error association algorithm, matches and optimizes the combination of the SAR image position and the AIS position, including: The global minimum mean square error association algorithm is used to calculate the minimum mean square value of the distance deviation of all ships in the longitude direction. The designed objective function is as follows: （6） in, Indicates that all the The average value of is a binary variable, 0 means no association, 1 means association, in a set of rough matching There is at most one associated value in Exceeds the number of matching pairs in this group When The value is 0; Indicates that all ships are roughly matched The maximum value of The Munkres algorithm is used to perform global minimum mean square error point matching calculation to obtain new matching point pairs between the SAR image position and the AIS position. ,in , represents the number of ships that achieved matching; calculate Euclidean distance between the SAR image position and AIS position of the vessel : （7） Calculate all matching point pairs The mean and standard deviation , judge whether the associated matching point pair exceeds the threshold. If it exceeds the threshold, it means that the associated matching is invalid and the matching point is deleted. The threshold range is , and obtain the matching point pairs after spatiotemporal correlation optimization .

8. The method according to claim 7, characterized in that In step S17, based on the optimized matching point pairs and the orbital position of the spaceborne SAR satellite corresponding to the actual imaging time of each ship, the AIS position is used as a control point to correct the RD model error, and the new geographic coordinate information of each SAR image pixel value is obtained again to obtain a geometrically corrected image product with improved positioning accuracy, including: The slope distance equation of each pixel in the RD model is expressed as the following formula (8): in, Indicates the initial slant distance used in slant distance calculation and the near distance value in imaging processing, which is obtained from the metadata of the current foreground image product; is the speed of light constant, is the number of sampling points in the distance direction; Indicates the sampling frequency of the echo signal, which is obtained from the metadata of the current foreground image product; Represents the range error, including the delay error introduced by atmospheric propagation, the system sampling time error and the channel delay error. This error value takes a fixed value in a single imaging task; The azimuth starting time error formula is as follows: （9） in, Indicates the starting time of the orientation in the imaging process, which is obtained from the metadata of the current foreground image product; is the azimuth pulse count; Indicates the time error in azimuth; According to formula (3), the orbital position error of the spaceborne SAR satellite is Expressed as a first-order function of the azimuth time error: （10） Combining formula (8) and formula (10), we can get the initial slant range error in the range direction: and azimuth starting time error Equation (11): According to equation (11), calculate and The RD model is modified according to the value of 9. The method according to claim 8, characterized in that According to equation (11), calculate and The RD model is modified based on the value of Step S171, obtaining calculation parameters, the calculation parameters include: the longitude and latitude coordinates of the AIS are expressed as the coordinate position of the ship in the SAR image , corresponding to the column coordinates of the ship on the spaceborne SAR image , and the parameters in the metadata and , and the actual imaging time of the vessel The corresponding orbital position of the spaceborne SAR satellite ; Step S172: Use the matching point pairs optimized by time-space correlation ,calculate and The RD model is modified by express The number of matches in , including: when , using formula (8), substitute , , , and calculate , and take the average value as the range slant range error , default azimuth time error is 0; when , using formula (10), we can obtain equations, set the initial and are all 0, and the Newton iteration method is used to calculate using the first-order derivative and ; Determine whether the error is less than the preset threshold through iterative calculation ,when When , the iteration is terminated, otherwise, the process goes to step S171 to continue the iterative calculation; Step S173: and As a result, the RD model equations are updated, and the corrected system geometry correction algorithm equations include: （12） （13） （14） The geographic coordinate position information of each pixel in the image is obtained again through iterative solution or the Ferrari solution, and finally a system geometric correction image product with improved positioning accuracy is obtained.

10. A system for improving the geometric positioning accuracy of spaceborne SAR ocean images, characterized in that: include: The data acquisition and processing module is used to formulate the future imaging mission planning and data transmission reception mission of the spaceborne SAR satellite system, determine the satellite imaging area range and the specific imaging time according to the imaging mission planning, and collect AIS data within the corresponding time and area range through the AIS data service provider after the SAR satellite completes imaging according to the imaging mission planning; It is also used to pre-process the original code stream data received by the data transmission station of the spaceborne SAR satellite ground processing system according to the data transmission reception task to obtain the spaceborne SAR ocean image with different scenes; A calculation module, used for detecting ships on the satellite-borne SAR ocean image and calculating the actual imaging time of each detected ship, so as to synchronize the time when fusing with the AIS data; It is also used to calculate the orbital position of the spaceborne SAR satellite corresponding to the actual imaging time of each ship based on the GPS data and the actual imaging time of each ship in combination with the dynamic model; A matching module is used to pre-match the ship in the satellite-borne SAR ocean image and the AIS data according to the actual imaging time of each ship, so as to obtain a combination of the SAR image position and the AIS position of the ship; An optimization module, used for performing matching optimization on the combination of the SAR image position and the AIS position based on a global minimum mean square error to obtain an optimized matching point pair; The correction module is used to correct the RD model error based on the optimized matching point pairs and the orbital position of the space-borne SAR satellite corresponding to the actual imaging time of each ship, using the AIS position as the control point, and re-solve to obtain the new geographic coordinate information of each SAR image pixel value to obtain a geometrically corrected image product with improved positioning accuracy.

Images