Ocean-sea aerial aircraft target satellite-borne SAR detection imaging method based on guidance information
By combining guidance information with a spaceborne SAR detection method for distant-sea aircraft targets, and utilizing ADS-B signals and spaceborne SAR data, rapid and accurate detection and imaging of aircraft targets have been achieved. This solves the problem of imaging aircraft targets under adverse weather conditions using spaceborne SAR, and improves imaging clarity and detection rate.
Patent Information
- Application Number
- CN202511780895.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-28
- Publication Date
- 2026-02-24
AI Technical Summary
Existing technologies make it difficult for spaceborne SAR to effectively detect and image aerial aircraft targets, especially under conditions of low signal-to-noise ratio, low signal-to-clutter ratio, weak target RCS, and high target speed. It is difficult to successfully capture and focus the signal of aerial aircraft targets.
A spaceborne SAR detection method for aircraft targets in the open sea, which combines guidance information, is proposed. This method involves acquiring global aircraft ADS-B signals for preprocessing, constructing a database, using the time and location information of SAR images for coarse screening, combining the CA-CFAR algorithm and deep learning for target detection, and applying an autofocus method to achieve fine focusing.
It enables rapid and accurate detection and imaging of aircraft targets in the open sea, improves the ability of spaceborne SAR to monitor aerial targets under adverse weather conditions, and enhances the signal-to-clutter ratio and imaging clarity of the targets.
Smart Images
Figure CN121559516A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of Synthetic Aperture Radar (SAR) signal processing technology, specifically relating to a spaceborne SAR detection and imaging method for distant-sea and airborne aircraft targets based on guidance information. Background Technology
[0002] Airborne target monitoring plays a vital role in the world today, and in the civil aviation sector, it is crucial for ensuring flight safety. By monitoring air traffic, conflicts between aircraft can be avoided, reducing the occurrence of air traffic accidents. Simultaneously, monitoring systems can provide weather information, helping pilots avoid adverse weather conditions and ensuring flight safety. With globalization and rapid technological advancements, air traffic is becoming increasingly busy, making real-time monitoring and accurate observation of airborne targets particularly critical. This not only relates to the smooth and safe flow of air traffic but also involves the timely detection and response to potential threats.
[0003] With technological advancements, methods for monitoring aerial targets are constantly evolving. While traditional ground-based radar systems and optical observation systems can provide aerial target monitoring to some extent, they each have their limitations. For example, ground-based radar is restricted by the horizon, terrain, and national borders, limiting its reconnaissance range, while optical observation is constrained by weather and lighting conditions. Spaceborne Synthetic Aperture Radar (SAR), unaffected by climate and environment, can operate stably in fog, rain, snow, and dark nights, continuously acquiring surface information. This characteristic allows high-resolution spaceborne SAR to provide valuable observational data even under adverse weather conditions. [Lan G. Cumming, Frank H. Wong. Synthetic Aperture Radar Imaging: Algorithms and Implementation [M]. Electronic Industry Press, 2012.]. Therefore, synthetic aperture radar (SAR) technology, with its all-weather, all-time operation capabilities and high-resolution multi-mode imaging capabilities, has become an important supplement to aerial target monitoring.
[0004] Currently, there are no publicly available publications on experimental results of spaceborne SAR Air Moving Target Indication (AMTI). Research on SAR targeting high-altitude moving aircraft targets is severely lacking, which is closely related to the difficulties of SAR observation of high-altitude, high-speed targets. First, spaceborne SAR is a ground observation method. When photographing aerial targets, ground targets are considered background clutter. Airborne aircraft targets have a small RCS (Radar Cross Section), making them easily mixed into the background clutter and difficult to detect. Second, compared to ground targets, aerial aircraft targets move faster, maneuver more complexly, and have more freedom in their directional dimensions, making it difficult to accurately estimate their slant range history and trajectory, thus making focusing on aerial targets extremely difficult. Therefore, the main challenges of spaceborne SAR detection and imaging of aerial aircraft targets can be summarized as how to successfully capture and detect the signals of aerial aircraft targets and focus them for imaging under conditions of low signal-to-noise ratio, low signal-to-clutter ratio, weak target RCS, and high target speed.
[0005] In summary, given the challenges of spaceborne SAR aerial aircraft target detection and imaging technology, there is an urgent need to conduct targeted research on aerial aircraft target data acquisition, target detection, and fine-focused imaging of moving targets. This research aims to fill the gap in spaceborne SAR aerial aircraft target detection and imaging research and solve the problem of the usability of SAR for aerial aircraft target identification. Summary of the Invention
[0006] The purpose of this invention is to provide a spaceborne SAR detection and imaging method for distant-sea and airborne aircraft targets based on guidance information.
[0007] The technical solution to achieve the purpose of this invention is: a spaceborne SAR detection and imaging method for long-range airborne aircraft targets based on guidance information, comprising the following steps:
[0008] Step 1: Obtain global aircraft ADS-B signals from the FlightAware website, perform preprocessing operations, including data cleaning and standardization, handling outliers and missing values, and noise interference processing, and build a global maritime aircraft information database.
[0009] Step 2: Obtain the SAR image acquisition time period and the latitude and longitude information of the four corner points of the image from the spaceborne SAR product database;
[0010] Step 3: Compare the latitude, longitude and timestamp information of the aircraft in the Global Maritime Aircraft Information Database with the time period of the SAR image and the latitude and longitude information of the four corner points of the image in the SAR Product Information Database to obtain SAR images that may contain aircraft targets.
[0011] Step 4: For SAR images that may contain aircraft targets, obtain L1A level single-view complex image data of the scene data by querying the scene number. Decompose the velocity and heading provided by the ADS-B of the suspected aircraft target captured by SAR according to the trajectory captured by SAR, and use the target azimuth velocity to achieve coarse focusing of the aircraft target.
[0012] Step 5: Apply the CA-CFAR algorithm to the coarse focusing imaging results to perform target detection and obtain a detection slice of the aircraft target.
[0013] Step 6: Obtain the refocused aerial moving target using an autofocus method.
[0014] Further, step 1: Obtain global aircraft ADS-B signals from the FlightAware website. The ADS-B signal contains: the ICAO code that uniquely identifies each aircraft, the longitude and latitude of the aircraft, flight altitude, flight speed, flight heading, and the specific time when the data was recorded.
[0015] Further, step 3: Compare the latitude, longitude, and timestamp information of the aircraft in the global maritime aircraft information database with the SAR image capture time period and the latitude and longitude information of the four corner points of the image in the SAR product information database to obtain a rough screening of spaceborne SAR aerial aircraft target images. The specific method is as follows:
[0016] Suppose the information of a certain aircraft in ADS-B is (longitude) ,latitude timestamp The time period for capturing SAR images is The latitude and longitude of the four corner points are at the top left. Bottom left corner Top right corner bottom right corner A cross-sectional analysis was conducted, with the following conditions and rules:
[0017]
[0018] If the conditions are met, it indicates that there may be an aircraft target in the SAR image. Record the aircraft's flight speed and direction information, and also record the scene number information of the SAR image.
[0019] Further, step 4: For SAR images that may contain aircraft targets, obtain L1A level single-look complex image data of the scene data by querying the scene number. Decompose the velocity and heading provided by the ADS-B of the suspected SAR-captured aircraft target according to the SAR-captured trajectory. Use the target's azimuth velocity to achieve coarse focusing of the aircraft target. Specifically, using the known target azimuth velocity can achieve coarse focusing of the aircraft target.
[0020] The aircraft velocity is decomposed into azimuth velocity along the SAR trajectory. and distance velocity According to the principles of SAR imaging and the theory of moving target analysis, the target's azimuth velocity will cause a change in the target's Doppler modulation frequency, specifically:
[0021]
[0022] in This indicates the change in Doppler modulation frequency caused by the target's azimuth velocity. This indicates the equivalent velocity used during imaging. Indicates wavelength. The reference distance is indicated, and all of the above parameters can be obtained from the SAR product metadata file;
[0023] Its corresponding azimuth error phase is:
[0024]
[0025] Where t represents the synthetic aperture time, which is obtained from the SAR product metafile;
[0026] Set the single-view complex image in the SARL1A product to Where x represents azimuth information and y represents range information, the azimuth frequency domain data is obtained by performing an azimuth Fourier transform. :
[0027]
[0028] in Indicates azimuth frequency information. The Fourier transform, along the azimuth direction, compensates for the phase error in the above equation to obtain:
[0029]
[0030] Performing an inverse azimuth Fourier transform on the above equation yields the coarse-focused imaging result of the moving target:
[0031] .
[0032] Further, in step 5, the CA-CFAR algorithm is applied to the coarse focusing imaging results for target detection to obtain a detection slice of the aircraft target. The specific method is as follows:
[0033] (1) Determine the detection window: Select a preset number of reference units around the unit to be detected to estimate the background noise power; at the same time, set up a protection unit to shield the influence of strong target echoes on noise estimation;
[0034] (2) Estimating background noise: Statistically process the amplitude of the reference unit selected in (1) and calculate the average value. This value is the estimated value of the background noise of the unit to be detected.
[0035] (3) Calculate the detection threshold: Based on the required constant false alarm probability and combined with the preset threshold factor (scaling factor), the background noise estimate obtained in (2) is amplified according to the threshold factor to obtain the detection threshold of the unit to be detected;
[0036] (4) Execute detection decision: Compare the amplitude of the unit to be detected with the detection threshold obtained in (3); if the unit to be detected is greater than the threshold, it is determined that there is a target; otherwise, it is determined that there is no target.
[0037] (5) Threshold adaptive adjustment: Under different clutter or noise environments, the threshold factor is adjusted online according to the detection output and the predetermined constant false alarm probability to maintain the detection rate.
[0038] Further, in step 5, a deep learning-based detection algorithm is used to perform target detection and obtain a detection slice of the aircraft target.
[0039] Furthermore, in step 6, the refocused aerial moving target is obtained through an autofocus method, specifically as follows:
[0040] (1) Based on the obtained target detection slice results, the data is transformed to the Doppler domain by inverse multiplication using an azimuth matched filter, resulting in the uncompressed azimuth data:
[0041]
[0042] Where c represents the speed of light. The value represents the target scattering intensity, n represents the range sampling point, m represents the azimuth sampling point, and ΔR(m) and Δφ(m) represent the range migration and phase error caused by the target motion, respectively.
[0043] (2) In view of the narrow frequency band of the air target and the serious frequency band offset, the spectrum is first circularly shifted to ensure the integrity of the target frequency band. Then the target frequency band range is extracted, and the minimum entropy range alignment algorithm is used for range alignment processing for this part of the frequency band.
[0044] Let the obtained distance alignment curve be Where m represents the azimuth sampling point, and the unit of the curve is pixels, it is converted into a range-time representation according to the relationship between pixels and range-time:
[0045]
[0046] in This represents the range alignment sampling rate. A Hampel filter is applied, utilizing a median-based nonlinear filtering method to identify and remove outliers in the range alignment curve while preserving its basic shape. A secondary fitting is then performed to obtain the processed range alignment curve. ;
[0047] After applying filtering to remove outliers, a second fitting is performed, and finally, the data is compensated and added to the azimuth frequency domain data.
[0048]
[0049] at this time This item should be adjusted to zero;
[0050] (3) After the distance is aligned, the phase autofocusing algorithm, the phase error correction algorithm based on the optimal image contrast, or the phase error correction algorithm based on the optimal minimum entropy of the image are applied to the data for phase autofocusing.
[0051] A satellite-borne SAR detection and imaging system for long-range airborne aircraft targets based on guidance information is proposed. This system implements the aforementioned satellite-borne SAR detection and imaging method for long-range airborne aircraft targets based on guidance information, achieving satellite-borne SAR detection and imaging of long-range airborne aircraft targets based on guidance information. The system is executed in six modules:
[0052] Module 1 retrieves global aircraft ADS-B signals from the FlightAware website, performs preprocessing operations including data cleaning and standardization, outlier and missing value handling, and noise interference processing, and builds a global maritime aircraft information database.
[0053] Module 2 retrieves the SAR image capture time period and the latitude and longitude information of the four corner points of the image from the spaceborne SAR product database;
[0054] Module 3: Compare the latitude, longitude, and timestamp information of aircraft in the Global Maritime Aircraft Information Database with the SAR image capture time period and latitude and longitude information of the four corner points of the image in the SAR Product Information Database to obtain SAR images that may contain aircraft targets.
[0055] Module 4: For SAR images that may contain aircraft targets, obtain L1A level single-view complex image data of the scene data by querying the scene number. Decompose the velocity and heading provided by the ADS-B of the suspected aircraft target captured by SAR according to the trajectory captured by SAR, and use the target azimuth velocity to achieve coarse focusing of the aircraft target.
[0056] Module 5 applies the CA-CFAR algorithm to the coarse focusing imaging results for target detection, and obtains a detection slice of the aircraft target;
[0057] Module 6 uses an autofocus method to refocus on moving aerial targets.
[0058] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the aforementioned method for spaceborne SAR detection and imaging of aircraft targets in the open sea based on guidance information, thereby realizing spaceborne SAR detection and imaging of aircraft targets in the open sea based on guidance information.
[0059] A computer-readable storage medium storing a computer program thereon, wherein when the computer program is executed by a processor, the aforementioned method for satellite-borne SAR detection and imaging of aircraft targets in the open sea based on guidance information is implemented, thereby realizing satellite-borne SAR detection and imaging of aircraft targets in the open sea based on guidance information.
[0060] Compared with existing technologies, the significant advantages of this invention are as follows: This invention proposes a rapid processing technology for integrated detection and imaging of long-range maritime and aerial aircraft targets, combining guidance information. Existing technologies do not possess a method entirely consistent with the objectives of this invention, and currently, no experimental results on spaceborne SAR aerial moving target detection and imaging have been published in the literature. This invention is forward-looking and innovative. Firstly, it innovatively proposes a spaceborne SAR method for detecting long-range maritime and aerial aircraft targets by combining publicly available ADS-B guidance information. Secondly, it utilizes this method to achieve detection and fine-focused imaging of long-range maritime and aerial aircraft targets using spaceborne SAR. Attached Figure Description
[0061] Figure 1 This section describes the global division of offshore areas.
[0062] Figure 2 The flowchart shows the minimum entropy distance alignment algorithm.
[0063] Figure 3 This is a flowchart of the minimum entropy phase correction algorithm.
[0064] Figure 4 This is a basic flowchart of the present invention.
[0065] Figure 5 This is a schematic diagram illustrating a case where an aircraft flight path was detected to coincide with the location of a SAR image at the same time.
[0066] Figure 6 This is a schematic diagram showing an aircraft being detected in GF3 data for a certain scene.
[0067] Figure 7 The result is a finely focused image of the aircraft target.
[0068] Figure 8 This is a fine-focused result of SAR aircraft data from the GF3 satellite. Detailed Implementation
[0069] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.
[0070] This invention discloses a spaceborne SAR detection and imaging method for distant-sea aerial aircraft targets based on guidance information. This method achieves coarse screening of moving aerial targets by combining optimized ADS-B (Automatic Dependent Surveillance-Broadcast) guidance information with spaceborne SAR target information; performs aerial target detection in the image domain using the CFAR (Constant False Alarm Rate) detection method; corrects the range migration of moving targets using a range alignment algorithm based on a portion of the spectrum; and achieves fine focusing of aerial targets using a sparse-combined minimum entropy autofocus algorithm. The specific steps are as follows:
[0071] Step 1: Generation of a real-time global aircraft information database
[0072] Global aircraft ADS-B signals refer to signals emitted by the Automatic Dependent Surveillance-Broadcast (ADS-B) system. The information contained primarily includes: the International Civil Aviation Organization (ICAO) code uniquely identifying each aircraft, the aircraft's longitude and latitude, altitude, speed, heading, and the specific time the data was recorded. This information is available in real-time on the publicly accessible website FlightAware (https: / / zh.flightaware.com).
[0073] This invention targets aircraft targets over the open sea. Firstly, the background over the open sea is clean with low clutter energy, making it easier to detect and image aircraft targets. Secondly, the most frequent flight paths for aircraft around the world are over the sea, making it more likely to capture aircraft targets over the open sea.
[0074] The execution program of this invention obtains aircraft information on global ocean locations from the FlightAware website. For efficiency, this invention divides the global ocean region into 27 areas, as detailed below. Figure 1 As shown. This area division is related to flight routes and aircraft density, and was obtained based on the highest efficiency after multiple multi-day experiments.
[0075] The aircraft information obtained from the global ocean locations is compiled into a global real-time aircraft information database.
[0076] Step 2: ADS-B signal preprocessing
[0077] The global aircraft information real-time database contains raw data of ADS-B signals of aircraft worldwide within a certain region over a specific period. This raw data typically contains various noises and inconsistencies. The method of this invention first preprocesses all information in the raw data, including the aircraft's longitude and latitude, flight altitude, flight speed, and flight heading. The specific steps are as follows:
[0078] (1) Data cleaning and standardization: Data standardization is achieved through format unification, unit conversion and deduplication;
[0079] (2) Outlier and missing value processing: For the same aircraft, flight information at different times is obtained. The outlier of the target data is determined by the degree of deviation of numerical fitting. The detected outlier is fitted and smoothed. Usually, the altitude, speed and heading of the aircraft at high altitude have a certain stability, so second-order fitting is sufficient. Finally, missing data is filled or deleted.
[0080] (3) Noise interference processing: To address the noise interference problem in ADS-B signals, noise separation of ADS-B signals is performed based on the FastICA (Fast Independent Component Analysis) algorithm to improve signal quality.
[0081] The preprocessed ADS-B signals are compiled into a global real-time preprocessed aircraft information database.
[0082] Step 3: Acquisition of spaceborne SAR product data
[0083] This step requires users of this invention to have access to publicly available or internal spaceborne SAR product data and their corresponding product information. For ease of explanation later, this part will be referred to as the spaceborne SAR product database, which generally includes spaceborne SAR L1A, L1B, and L2 level products, as well as corresponding metadata. Each SAR image also has a unique number that can be used to query its corresponding L1A, L1B, and L2 level products. The example shown in this invention comes from the China Ocean Satellite Data Service System website (https: / / osdds.nsoas.org.cn / home), where on-orbit operational product data information for the GF03 SAR satellite can be obtained.
[0084] The present invention provides an execution program module that reads the metadata file of L2-level SAR satellite products, obtains the unique number, shooting time period, and latitude and longitude information of the four corner points of all SAR images in the spaceborne SAR product database, and organizes the above information into a spaceborne SAR time and location database for later use.
[0085] Step 4: Coarse screening of aircraft targets in the open sea and air by spaceborne SAR
[0086] By executing the program of the long-range airborne aircraft target coarse screening module of the present invention, and by comparing the latitude and longitude and timestamp information of the aircraft in the global aircraft information real-time preprocessing database obtained in step 2 with the SAR image shooting time period and latitude and longitude information of the four corner points of the image in the spaceborne SAR time and location database obtained in step 3, a coarse screening of spaceborne SAR airborne aircraft target images can be obtained.
[0087] Suppose the information of an aircraft in ADS-B is (longitude) ,latitude timestamp The time period for capturing a certain spaceborne SAR image is as follows: The latitude and longitude of the four corner points are at the top left. Bottom left corner Top right corner bottom right corner By cross-analyzing the above information, if the following conditions are met, it indicates that there may be captured aircraft targets in the SAR data of that scene.
[0088]
[0089] At this point, record the aircraft's flight speed and direction information, and also record the unique identifier of the SAR data for that scene.
[0090] Step 5: Coarse focusing of aerial moving targets based on guidance information
[0091] For offshore SAR data that may contain aircraft targets, the L1A-level product of the scene can be obtained by querying the spaceborne SAR product database using a unique number.
[0092] The speed and heading data (obtained from the global aircraft information real-time preprocessing database) of the suspected SAR-captured aircraft targets obtained in step 4 are decomposed according to the trajectory captured by the satellite-borne SAR. The aircraft speed is decomposed into azimuth speed along the SAR trajectory direction. and distance velocity Azimuth velocity can cause defocusing of moving targets, reducing their energy. Therefore, aircraft targets that may be present in the initial SAR image will be defocused, making subsequent detection more difficult.
[0093] Using the known target azimuth velocity, coarse focusing of the aircraft target can be achieved. According to the principles of SAR imaging and moving target analysis theory, the target azimuth velocity will cause a change in the target's Doppler modulation frequency, with the specific value being approximately:
[0094]
[0095] in This indicates the change in Doppler modulation frequency caused by the target's azimuth velocity. This indicates the equivalent velocity used during imaging. Indicates wavelength. The reference distance is indicated. All of the above parameters can be obtained from the L1A-level product metadata file in the spaceborne SAR product database.
[0096] Its corresponding azimuth error phase is:
[0097]
[0098] Where t represents the synthetic aperture time, which can be obtained from the SAR product metafile.
[0099] Set the single-view complex image in the SARL1A product to Where x represents azimuth information and y represents range information. An azimuth Fourier transform is performed to obtain the azimuth frequency domain data. :
[0100]
[0101] in Indicates azimuth frequency information. The Fourier transform of this data, along the azimuth direction, compensates for the phase error in the above equation, yielding:
[0102]
[0103] Performing an inverse azimuth Fourier transform on the above equation yields the coarse-focused imaging result of the moving target:
[0104]
[0105] Step 6: Aerial Target Detection
[0106] The Cell Averaging CFAR (CA-CFAR) algorithm is applied to target detection in the coarse focusing imaging results. The specific steps are as follows:
[0107] (1) Determine the detection window: Select a preset number of reference units around the unit to be detected to estimate the background noise power; at the same time, set up a protection unit to shield the influence of strong target echoes on noise estimation;
[0108] (2) Estimating background noise: Statistically process the amplitude of the reference unit selected in (1) and calculate the average value. This value is the estimated value of the background noise of the unit to be detected.
[0109] (3) Calculate the detection threshold: Based on the required constant false alarm probability and combined with the preset threshold factor (scaling factor), the background noise estimate obtained in (2) is amplified according to the threshold factor to obtain the detection threshold of the unit to be detected;
[0110] (4) Execute detection decision: Compare the amplitude of the unit to be detected with the detection threshold obtained in (3); if the unit to be detected is greater than the threshold, it is determined that there is a target; otherwise, it is determined that there is no target.
[0111] (5) Threshold adaptive adjustment: Under different clutter or noise environments, the threshold factor is adjusted online according to the detection output and the predetermined constant false alarm probability to maintain the detection rate.
[0112] If deep learning-based detection algorithms or other improved CFAR detection algorithms are used, they also fall within the scope of protection of this invention patent.
[0113] Step 7: Fine-focusing imaging
[0114] The aircraft target has a complex movement pattern and a high speed. In addition to defocusing caused by phase error, there is also a serious distance migration phenomenon. The coarse focusing in step 5 is to uniformly compensate for the whole image. The compensation accuracy is limited by the accuracy of prior knowledge and cannot handle the cross-range gate defocusing caused by distance migration. Therefore, after obtaining the detection slice data of the aircraft target, the present invention continues to perform fine focusing imaging on the aircraft target slice.
[0115] The purpose of the fine-focusing method is to obtain a refocused aerial moving target through an autofocusing method, thereby improving the target's signal-to-clutter ratio. The specific steps are as follows:
[0116] (1) Azimuth Fourier Transform
[0117] Based on the target detection slice results obtained in step 6, fine focusing performs further focusing processing on target details. Therefore, it is necessary to transform the data to the Doppler domain by inverse multiplication using an azimuth matched filter, thus obtaining the uncompressed azimuth data:
[0118]
[0119] Where c represents the speed of light. The value represents the target scattering intensity, n represents the range sampling point, m represents the azimuth sampling point, and ΔR(m) and Δφ(m) represent the range migration and phase error caused by the target motion, respectively.
[0120] (2) Distance migration correction based on partial spectrum
[0121] Airborne targets move at high speeds. The synthetic aperture time of SAR is generally on the order of seconds, and the swath width is on the order of kilometers. Most of the time, high-speed moving targets in the air cannot be illuminated by the full aperture. Therefore, their spectrum only occupies a small part of the full frequency band. The rest is equivalent to background clutter for airborne targets. At the same time, because the radial velocity of the target is too fast, the Doppler spectrum will also be significantly shifted, and may even exceed the azimuth sampling rate.
[0122] Traditional range alignment algorithms are highly dependent on the data signal-to-noise ratio, which can cause range alignment algorithms based on full-band data to fail. Doppler spectral shift also affects range alignment algorithms to some extent. To address the narrow bandwidth and severe bandwidth shift of airborne targets, this paper first performs a circular shift of the spectrum to ensure the integrity of the target bandwidth. Then, the target bandwidth range is extracted, and range alignment processing is performed on this portion of the bandwidth using a minimum entropy range alignment algorithm. The flowchart of this algorithm is shown below. Figure 2 As shown.
[0123] Let the obtained distance alignment curve be Where m represents the azimuth sampling point, and the unit of the curve is pixels, it can be converted into range-time according to the relationship between pixels and range-time as follows:
[0124]
[0125] in This represents the range alignment sampling rate. A Hampel filter is applied, utilizing a median-based nonlinear filtering method to identify and remove outliers in the range alignment curve while preserving its basic shape. A secondary fitting is then performed to obtain the processed range alignment curve. .
[0126] After applying filtering to remove outliers, a second fitting is performed, and finally, the data is compensated and added to the azimuth frequency domain data.
[0127]
[0128] at this time This item should be basically corrected to zero.
[0129] (3) Phase compensation and focusing
[0130] After distance alignment, the data is processed by applying the minimum entropy phase correction algorithm for phase autofocusing, which can compensate for the △φ(m) phase term and make the image focus.
[0131] If a phase gradient autofocus algorithm or a phase error correction algorithm based on optimal image contrast is used, it also falls within the protection scope of this invention patent.
[0132] Example
[0133] This invention innovatively proposes a spaceborne SAR method for detecting distant maritime and aerial aircraft targets by combining publicly available ADS-B guidance information. Furthermore, it utilizes this method to achieve spaceborne SAR detection and fine-focused imaging of distant maritime and aerial aircraft targets. The research results of this invention are illustrated below using GF3 spaceborne SAR data as an example.
[0134] After the initial screening of distant aerial targets in steps 1 to 4, data from a specific scene in GF3 fine strip mode overlapped with the flight path of a certain aircraft during the shooting time. Figure 5 The flight path extracted by ADS-B was displayed. Figure 6 The image shows a secondary HV polarization product of the GF3 data for this scene. The parameters of this data are shown in Table 1. An aircraft target in flight was detected in the upper right corner.
[0135] Table 1. GF3 data parameters for a certain scene
[0136]
[0137] By focusing precisely on the aircraft target, we can obtain... Figure 7 The processing comparison results show that... Figure 7 The aircraft's outline is clearly visible in (b). Furthermore, this invention has also detected aerial aircraft targets in other GF3 data; the fine-focusing results are listed below. Figure 8 As can be seen from the image, the aircraft target has been finely processed with a significantly improved focus, and its outline is clearly visible. This invention demonstrates the effectiveness of the technology.
[0138] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0139] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these modifications and improvements all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.
Claims
1. A spaceborne SAR detection and imaging method for long-range aerial aircraft targets based on guidance information, characterized in that, Includes the following steps: Step 1: Obtain global aircraft ADS-B signals from the FlightAware website, perform preprocessing operations, including data cleaning and standardization, handling outliers and missing values, and noise interference processing, and build a global maritime aircraft information database. Step 2: Obtain the SAR image acquisition time period and the latitude and longitude information of the four corner points of the image from the spaceborne SAR product database; Step 3: Compare the latitude, longitude and timestamp information of the aircraft in the Global Maritime Aircraft Information Database with the time period of the SAR image and the latitude and longitude information of the four corner points of the image in the SAR Product Information Database to obtain SAR images that may contain aircraft targets. Step 4: For SAR images that may contain aircraft targets, obtain L1A level single-view complex image data of the scene data by querying the scene number. Decompose the velocity and heading provided by the ADS-B of the suspected aircraft target captured by SAR according to the trajectory captured by SAR, and use the target azimuth velocity to achieve coarse focusing of the aircraft target. Step 5: Apply the CA-CFAR algorithm to the coarse focusing imaging results to perform target detection and obtain a detection slice of the aircraft target. Step 6: Obtain the refocused aerial moving target using an autofocus method.
2. The method for spaceborne SAR detection and imaging of long-range airborne aircraft targets based on guidance information according to claim 1, characterized in that, Step 1: Obtain global aircraft ADS-B signals from the FlightAware website. The ADS-B signal contains: the ICAO code that uniquely identifies each aircraft, the longitude and latitude of the aircraft, flight altitude, flight speed, flight heading, and the specific time when the data was recorded.
3. The method for spaceborne SAR detection and imaging of long-range airborne aircraft targets based on guidance information according to claim 1, characterized in that, Step 3: Compare the latitude, longitude, and timestamp information of the aircraft in the global maritime aircraft information database with the SAR image capture time period and the latitude and longitude of the four corner points of the image in the SAR product information database to obtain a rough screening of spaceborne SAR aerial aircraft target images. The specific method is as follows: Suppose the information of a certain aircraft in ADS-B is (longitude) ,latitude timestamp The time period for capturing SAR images is The latitude and longitude of the four corner points are at the top left. Bottom left corner Top right corner bottom right corner A cross-sectional analysis was conducted, with the following conditions and rules: If the conditions are met, it indicates that there may be an aircraft target in the SAR image. Record the aircraft's flight speed and direction information, and also record the scene number information of the SAR image.
4. The method for spaceborne SAR detection and imaging of long-range airborne aircraft targets based on guidance information according to claim 1, characterized in that, Step 4: For SAR images that may contain aircraft targets, obtain L1A level single-look complex image data of the scene data by querying the scene number. Decompose the velocity and heading provided by the ADS-B of the suspected SAR-captured aircraft target according to the SAR-captured trajectory. Use the target's azimuth velocity to achieve coarse focusing of the aircraft target. Specifically, using the known target azimuth velocity can achieve coarse focusing of the aircraft target. The aircraft velocity is decomposed into azimuth velocity along the SAR trajectory. and distance velocity According to the principles of SAR imaging and the theory of moving target analysis, the target's azimuth velocity will cause a change in the target's Doppler modulation frequency, specifically: in This indicates the change in Doppler modulation frequency caused by the target's azimuth velocity. This indicates the equivalent velocity used during imaging. Indicates wavelength. The reference distance is indicated, and all of the above parameters can be obtained from the SAR product metadata file; Its corresponding azimuth error phase is: Where t represents the synthetic aperture time, which is obtained from the SAR product metafile; Set the single-view complex image in the SARL1A product to Where x represents azimuth information and y represents range information, the azimuth frequency domain data is obtained by performing an azimuth Fourier transform. : in Indicates azimuth frequency information. The Fourier transform, along the azimuth direction, compensates for the phase error in the above equation to obtain: Performing an inverse azimuth Fourier transform on the above equation yields the coarse-focused imaging result of the moving target: 。 5. The method for spaceborne SAR detection and imaging of long-range airborne aircraft targets based on guidance information according to claim 1, characterized in that, Step 5: Apply the CA-CFAR algorithm to the coarse focusing imaging results for target detection to obtain a detection slice of the aircraft target. The specific method is as follows: (1) Determine the detection window: Select a preset number of reference units around the unit to be detected to estimate the background noise power; at the same time, set up a protection unit to shield the influence of strong target echoes on noise estimation; (2) Estimating background noise: Statistically process the amplitude of the reference unit selected in (1) and calculate the average value. This value is the estimated value of the background noise of the unit to be detected. (3) Calculate the detection threshold: Based on the required constant false alarm probability and combined with the preset threshold factor (scaling factor), the background noise estimate obtained in (2) is amplified according to the threshold factor to obtain the detection threshold of the unit to be detected; (4) Execute the detection decision: Compare the amplitude of the unit to be detected with the detection threshold obtained in (3); If the number of units to be detected is larger than the threshold, then a target is determined to exist; otherwise, no target is determined to exist. (5) Threshold adaptive adjustment: Under different clutter or noise environments, the threshold factor is adjusted online according to the detection output and the predetermined constant false alarm probability to maintain the detection rate.
6. The method for spaceborne SAR detection and imaging of long-range airborne aircraft targets based on guidance information according to claim 1, characterized in that, Step 5: Use deep learning-based detection algorithms to perform target detection and obtain detection slices of aircraft targets.
7. The method for spaceborne SAR detection and imaging of long-range airborne aircraft targets based on guidance information according to claim 1, characterized in that, Step 6: Obtain the refocused aerial moving target using an autofocus method. The specific method is as follows: (1) Based on the obtained target detection slice results, the data is transformed to the Doppler domain by inverse multiplication using an azimuth matched filter, resulting in the uncompressed azimuth data: Where c represents the speed of light. The value represents the target scattering intensity, n represents the range sampling point, m represents the azimuth sampling point, and ΔR(m) and Δφ(m) represent the range migration and phase error caused by the target motion, respectively. (2) In view of the narrow frequency band of the air target and the serious frequency band offset, the spectrum is first circularly shifted to ensure the integrity of the target frequency band. Then the target frequency band range is extracted, and the minimum entropy range alignment algorithm is used for range alignment processing for this part of the frequency band. Let the obtained distance alignment curve be Where m represents the azimuth sampling point, and the unit of the curve is pixels, it is converted into a range-time representation according to the relationship between pixels and range-time: in This represents the distance to time of the alignment curve. Indicates the distance sampling rate; By applying a Hampel filter and utilizing a median-based nonlinear filtering method, outliers in the range alignment curve are identified and removed while preserving its basic shape. A secondary fitting is then performed to obtain the processed range alignment curve. Finally, the compensation is added to the azimuth frequency domain data: at this time This item should be corrected to zero; (3) After the distance is aligned, the phase gradient autofocus algorithm, the phase error correction algorithm based on the optimal image contrast, or the phase error correction algorithm based on the optimal minimum entropy of the image are applied to perform phase autofocus processing.
8. A spaceborne SAR detection and imaging system for long-range aerial aircraft targets based on guidance information, characterized in that, The method for spaceborne SAR detection and imaging of aircraft targets in the open sea based on guidance information, as described in any one of claims 1-7, is implemented by executing six modules respectively: Module 1 retrieves global aircraft ADS-B signals from the FlightAware website, performs preprocessing operations including data cleaning and standardization, outlier and missing value handling, and noise interference processing, and builds a global maritime aircraft information database. Module 2 retrieves the SAR image capture time period and the latitude and longitude information of the four corner points of the image from the spaceborne SAR product database; Module 3: Compare the latitude, longitude, and timestamp information of aircraft in the Global Maritime Aircraft Information Database with the SAR image capture time period and latitude and longitude information of the four corner points of the image in the SAR Product Information Database to obtain SAR images that may contain aircraft targets. Module 4: For SAR images that may contain aircraft targets, obtain L1A level single-view complex image data of the scene data by querying the scene number. Decompose the velocity and heading provided by the ADS-B of the suspected aircraft target captured by SAR according to the trajectory captured by SAR, and use the target azimuth velocity to achieve coarse focusing of the aircraft target. Module 5 applies the CA-CFAR algorithm to the coarse focusing imaging results for target detection, obtaining a detection slice of the aircraft target; Module 6 uses an autofocus method to refocus on moving aerial targets.
9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements the spaceborne SAR detection and imaging method for distant-sea airborne aircraft targets based on guidance information as described in any one of claims 1-7, thereby realizing spaceborne SAR detection and imaging of distant-sea airborne aircraft targets based on guidance information.
10. A computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, it implements the spaceborne SAR detection and imaging method for distant-sea airborne aircraft targets based on guidance information as described in any one of claims 1-7, thereby realizing spaceborne SAR detection and imaging of distant-sea airborne aircraft targets based on guidance information.