High-temporal foundation SAR lunar rail interferometric processing method

Abstract

Interferometric synthetic aperture radar technology (InSAR) is an important technology in the field of synthetic aperture radar (SAR), which can realize the extraction of three-dimensional terrain. The present application provides a high-time-efficiency ground-based SAR lunar orbit interferometric processing method, which fills the gap of lunar interferometric SAR technology in this observation mode. Moreover, the method can realize lunar multi-baseline interferometry only by two observations, greatly reducing the time cost of multi-baseline interferometry.

Description

Technical Field

[0001] This invention relates to the field of radar signal processing, and in particular to a highly time-efficient ground-based SAR heavy orbit interferometry processing method for lunar surface topography extraction. Background Technology

[0002] Interferometric Synthetic Aperture Radar (InSAR) is an important technology in the field of Synthetic Aperture Radar (SAR), capable of extracting three-dimensional terrain. Using ground-based interferometric SAR for lunar observation can yield high-precision lunar surface topographic maps, providing crucial data support for studies on lunar origins, resource exploration, and spacecraft landing site selection, thus possessing significant scientific value.

[0003] Currently, both domestic and international applications of ground-based lunar interferometric SAR technology employ a single-transmitter / dual (or multiple)-receiver mode. This involves using one antenna to transmit a signal, while two (or more) antennas with a certain spatial baseline receive the echo. For example, the Haystack Radar system in the United States first used interferometric SAR technology to complete lunar observations in 1972, while the Goldstone Solar System Radar system used interferometric SAR technology to obtain lunar surface topographic maps, including the lunar poles, with a resolution of 5 meters, making it the benchmark for current lunar observation results.

[0004] Besides the single-transmitter, dual (or multiple)-receiver observation system, elevation inversion can also be achieved using a single antenna for re-orbit observations of the Moon, a technique known as re-orbit interferometric SAR. This mode requires two (or more) SAR imagings of the Moon at different times, extracting elevation by utilizing the interferometric baseline formed by the geometric differences in observations during ground-based radar revisits. Compared to the fixed baseline of the single-transmitter, dual (or multiple)-receiver mode, the re-orbit interferometric SAR mode can obtain more flexible and richer baselines to meet the needs of high-precision, complex terrain extraction. Furthermore, the re-orbit interferometric SAR technique requires only one antenna to complete terrain extraction, making it more cost-effective and economical.

[0005] However, the validation of lunar heavy orbit interferometric SAR technology remains incomplete, and a set of interferometric processing methods for this mode is currently lacking. Furthermore, and more importantly, lunar heavy orbit interferometric SAR has a significant drawback: extremely low timeliness. To meet the requirements of high-quality interferometric processing, heavy orbit interferometric SAR has strict length requirements for the interferometric baselines formed by multiple revisits; generally, shorter baselines are needed to effectively maintain spatial coherence. However, since the heavy orbit interferometric baselines are formed using the relative motion between the Earth and the Moon, the changes in the baselines are closely related to the period of this motion. Generally, an effective interferometry can only be formed after an integer multiple of the Moon's orbital period (approximately 27 days). In addition, for complex terrain areas, multi-baseline joint interferometric processing is often required, which can take several (or even a dozen) months to accumulate enough interferometric baselines to complete terrain inversion. Therefore, the time cost of lunar heavy orbit interferometric SAR processing is extremely high, and a high-efficiency lunar heavy orbit interferometric processing method is urgently needed to solve these problems. Summary of the Invention

[0006] In view of this, the present invention provides a high-efficiency ground-based SAR method for lunar multiple-orbit interferometry, filling the gap in lunar interferometric SAR technology under this observation mode. Furthermore, this method requires only two observations to achieve lunar multiple-baseline interferometry, greatly reducing the time cost of multiple-orbit interferometry.

[0007] The Earth-Moon relative motion model is attached. Figure 1 As shown, due to the Earth's rotation, ground-based radar has observable periods of time each day near the Moon. During different observable periods, the ground-based radar forms different tracks relative to the lunar surface. First, this invention filters track combinations based on coherence requirements to obtain track combinations that meet the needs of interferometric processing. Then, based on azimuth resolution requirements, azimuth aperture segmentation is performed on the master and slave tracks. Multiple sub-apertures are paired and combined to form multiple interferometric baselines, thereby maximizing the acquisition of the most interferometric baselines with the shortest re-orbit interval. Subsequently, based on the classic interferometric SAR processing workflow, the paired sub-aperture SAR images are registered, interferograms are extracted, and phase filtering is performed. Phase unwrapping is then performed based on the multi-baseline interferograms to meet the high-precision unwrapping requirements of complex terrain areas. Finally, elevation inversion is performed based on the unwrapped phases, and lunar surface DEM data is obtained through geocoding.

[0008] Beneficial effects:

[0009] This invention proposes for the first time a lunar interferometric SAR processing method in the double orbit mode, and adopts a sub-aperture segmentation method to form multi-baseline joint observations, which meets the requirements of high-precision terrain inversion in complex lunar terrain areas and greatly reduces the time cost of multi-baseline double orbit interferometric SAR.

[0010] On the one hand, the lunar heavy orbit interferometric SAR method proposed in this paper can complete the lunar surface elevation inversion with only a single antenna, which has low system construction cost and flexible and rich baselines, meeting the needs of complex terrain extraction.

[0011] On the other hand, this method uses sub-aperture segmentation to form multiple baselines, which has a significant time efficiency advantage compared to traditional multiple baseline formation methods using multiple orbits. Traditional methods require multiple orbit observations to form multiple baselines. Due to the periodicity of the relative motion between the Earth and the Moon, the time interval for a single baseline is generally an integer multiple of 27 days, while multiple baselines require several (or even a dozen) months. However, this method only requires two observations to form multiple baselines, drastically reducing time costs. Moreover, the more baselines required, the more obvious the advantage of this method becomes.

[0012] According to the present invention, by conducting two lunar observations with a single-antenna ground-based radar, multi-baseline joint interferometry processing of the moon can be achieved, effectively acquiring lunar surface DEM products and supporting lunar mapping tasks. Attached Figure Description

[0013] Figure 1 Earth-Moon Relative Motion Model Diagram

[0014] Figure 2 Flowchart of Ground-Based SAR Lunar Heavy Orbit Interferometry Processing Method

[0015] Figure 3 Schematic diagram of multi-baseline formation based on sub-aperture division

[0016] Figure 4 Geographic Coding Diagram Detailed Implementation

[0017] The present invention will now be described in detail with reference to the accompanying drawings.

[0018] This invention discloses a ground-based multiple orbit interferometric SAR processing method for lunar mapping. A flowchart of this method is attached. Figure 2 As shown, the process consists of two steps: Step 1 is the selection of the double-track interferometric baseline and the division of sub-apertures, corresponding to steps 1 and 2; Step 2 is the interferometric processing, corresponding to steps 3 to 7. Each step is described in detail below.

[0019] Step 1: Selection of Interferometric Baselines for Double Track Interference

[0020] Based on lunar ephemeris information, the relative motion trajectory between the Earth and the Moon can be obtained. The observed trajectories within a certain time window are grouped according to continuous visible periods (generally one day). Then, all groups are paired, and their vertical baselines are calculated. According to the requirements of spatial coherence, the vertical baselines within the critical baseline are filtered, and the corresponding observation period combinations are retained. The calculation of the critical baseline is shown in equation (1), where λ is the system wavelength, R is the distance between the radar and the observation center in the main trajectory, and θ... max The maximum elevation angle of the radar relative to the observation center in the main trajectory, ρ r η represents the distance resolution, and η represents the terrain slope angle.

[0021]

[0022] Step 2: Sub-aperture division

[0023] In ground-based double-orbit lunar interferometry, the relative motion between the Earth and the Moon during the interferometric observation period causes changes in the interferometric baseline. This characteristic can be used to divide the interferometric field into sub-apertures. After sub-aperture division, the corresponding sub-apertures between the master and slave observations form interferometric pairs, and different interferometric pairs have different line lengths, thus forming an effective multi-baseline interferometry.

[0024] During sub-aperture segmentation, to maintain sufficient terrain mesh refinement, the minimum aperture length needs to be determined based on the azimuth resolution requirements. The minimum aperture length determines the maximum number of corresponding sub-apertures, which are then used for subsequent segmentation.

[0025] Appendix Figure 3 The diagram illustrates sub-aperture division, with the master and slave trajectories corresponding to each other, separated by dashed lines. This division method allows for multiple interferometric pairs with different baseline lengths, enabling improved topographic measurement accuracy through multi-baseline joint processing. Furthermore, based on this method, multi-baseline interferometry processing can be completed with only two observations from a single radar system.

[0026] Step 3: SAR image registration and interferogram generation

[0027] In the process of interferometric SAR data processing, due to the certain degree of offset, stretching, and rotation of coherent pixels in two SAR images, images with the same coordinate grid do not correspond to the same ground scattering unit. To ensure that pixels at the same position in each SAR image correspond to the same ground scattering unit, SAR image registration is required. SAR image registration includes the registration between SAR images generated by the sub-apertures of the master and slave orbits, as well as the registration between master and slave SAR images. This invention uses the real correlation function method for registration.

[0028] After SAR image registration, the phase information is obtained by performing complex conjugate multiplication on the master and slave SAR images generated by each sub-aperture, which yields several interferometric fringe patterns with different baseline lengths.

[0029] Step 4: Go to the flatland phase

[0030] The flat-ground phase refers to the periodic phase formed by flat terrain with constant surface height, given a fixed orbital data. This phase needs to be removed when retrieving the true surface elevation. This invention calculates the slant range difference between pixels in master-slave SAR images based on orbital data, thereby calculating the phase difference caused by the flat terrain. Multiplying the calculated flat-ground phase by the interferogram removes the flat-ground phase.

[0031] Step 5: Phase Filtering

[0032] Due to the influence of various decoherence sources, the acquired lunar surface interferometric phase map contains significant noise, which restricts the efficiency and accuracy of subsequent phase unwrapping processing and needs to be filtered out. Because the interferometric phase exhibits periodic changes, with phase jumps at -π and π, general filtering methods cannot be used to directly denoise the phase. This invention uses the Goldstein filtering algorithm to process the lunar surface interferogram. The algorithm's principle is to divide the interferogram into overlapping sliding windows, calculate the power spectrum of the phase in each sliding window using Fourier transform, and then smooth the power spectrum to achieve the final filtering effect.

[0033] Step 6: Multi-baseline phase unwrapping

[0034] Because the present invention divides the sub-apertures during the baseline formation stage, it can obtain several interferograms with different baseline lengths. Therefore, the multi-baseline phase unwrapping technique can be used to effectively extract the steep terrain of the lunar surface, breaking through the limitation of phase stability in single-baseline phase unwrapping.

[0035] The core principle of the multi-baseline maximum likelihood estimation phase unwrapping algorithm lies in the existence of a probability density function for the interferometric phase as shown in equation (4):

[0036]

[0037] In the formula, φ and Let be the true interference phase and the unwrapped phase (or an estimate of the true phase), respectively, and γ be the coherence coefficient. This probability density function is periodic with a period of 2π. From the above equation, it can be seen that when the unwrapped phase... The probability density function reaches its maximum value when the phase is an accurate estimate of the true phase or when the phase difference is an integer multiple of 2π. Therefore, phase estimation can be performed by detecting the maximum value of the probability density function. Thus, multiplying the probability densities of all baselines yields the joint probability density function of the multiple baselines, expressed as:

[0038]

[0039] The process of multi-baseline maximum likelihood phase unwrapping is to solve for the estimated value of the interference phase corresponding to the maximum value of the joint probability density function in the non-fuzzy interval, which is expressed as Equation (6), where the subscript p represents different pixel points.

[0040]

[0041] Step 7: Elevation Inversion and Geographic Coding

[0042] After obtaining the true phase through multi-baseline unwrapping, the target elevation data can be retrieved by using the elevation ambiguity through phase inversion. In the heavy orbit InSAR system, the elevation ambiguity is expressed as shown in equation (7), which reflects the elevation change with each 2π phase change. Therefore, the target elevation can be calculated using equation (8), where... The interference phase is obtained after multi-baseline unwrapping.

[0043]

[0044]

[0045] The retrieved elevation lies within the azimuth-distance plane. Geocoding is used to transfer the elevation information to a spatial coordinate system, and then to geographic coordinates. A diagram of geocoding is attached. Figure 4 As shown.

Claims

1. A high-precision ground-based SAR lunar orbiting rail interferometric processing method, which utilizes a single ground-based radar to perform two lunar orbiting observations along the rail, and is based on the intersection of the master and slave observation tracks within the interferometric observation period due to the relative motion between the Earth and the Moon, and the time-varying baseline, characterized in that, Includes the following steps: Step 1: Baseline Screening for Double Orbit Interference: Based on lunar ephemeris information, the relative motion trajectory between the Earth and the Moon is obtained. The observed trajectories within a certain time window are grouped according to continuous visible periods. All groups are paired and their vertical baselines are calculated. Vertical baselines within the critical baseline are screened according to spatial coherence requirements, retaining the corresponding observation period combinations. The formula for calculating the critical baseline is: ； In the formula, For the system wavelength, The distance between the radar and the observation center in the main trajectory. The maximum elevation angle of the radar relative to the observation center in the main trajectory. For range resolution, The slope angle of the terrain; Step 2, Sub-aperture segmentation: Within the two master-slave observation tracks obtained through screening, based on the time-varying characteristics of the interferometric baseline caused by the relative motion of the Earth and the Moon, and according to the azimuth resolution requirements, the minimum sub-aperture length and the corresponding number of sub-apertures are determined. The master and slave tracks are segmented by azimuth aperture, so that the corresponding sub-apertures between the master and slave tracks are paired with each other to form multiple interferometric pairs with different baseline lengths, so as to improve the accuracy of topographic measurement by using multi-baseline joint processing. Step 3: SAR image registration and interferogram generation: The real correlation function method is used to register the SAR images generated by the sub-apertures of the master and slave orbits. The phase information is obtained by performing complex conjugate multiplication on the master and slave SAR images generated by each sub-aperture, resulting in several interferometric fringe patterns with different baseline lengths. Step 4: Remove the flat-ground phase: Calculate the slant range difference between pixels in the master-slave SAR images based on the master-slave orbit data, thereby calculating the phase difference caused by the flat-ground. Multiply the calculated flat-ground phase by the interferogram to remove the flat-ground phase. Step 5, Phase Filtering: The Goldstein filtering algorithm is used to process the lunar interferogram. The interferogram is divided into overlapping sliding windows. The power spectrum of the phase of each sliding window is calculated by Fourier transform and then smoothed to achieve the filtering effect. Step 6, Multi-baseline phase unwrapping: Based on the multi-baseline interferometric pair formed in Step 2, the multi-baseline maximum likelihood estimation phase unwrapping algorithm is used to construct the joint probability density function of all baselines. The estimated value of the interferometric phase corresponding to the maximum value of the joint probability density function is solved in the non-ambiguous interval, thereby realizing the multi-baseline joint phase unwrapping. Step 7, Elevation Inversion and Geocoding: After unwrapping multiple baselines, the true phase is obtained. The target elevation data is inverted using the elevation ambiguity, and the elevation information is transferred from the azimuth-distance plane to the lunar surface geographic coordinate system through geocoding to obtain the lunar surface DEM data.

2. The high-efficiency ground-based SAR lunar multiple orbit interferometry processing method as described in claim 1, characterized in that, The probability density function of the interferometric phase in the multi-baseline maximum likelihood estimation phase unwrapping algorithm in step six is: ； in, This is the true interference phase; This is the untangled phase, or an estimate of the true phase; The coherence coefficient is the probability density function, which is a periodic function with a period of . When the unwrapped phase is an accurate estimate or difference of the true phase. The probability density function reaches its maximum value when the probability density function is an integer multiple; phase estimation can be performed by detecting the maximum value of the probability density function; therefore, multiplying the probability densities of all baselines yields the joint probability density function of the multiple baselines, expressed as: ； The process of multi-baseline maximum likelihood phase unwrapping involves finding the estimated value of the interferometric phase in the unambiguous interval that maximizes the joint probability density function, expressed as: ； Subscript Represents different pixels.

3. The high-efficiency ground-based SAR lunar multiple orbit interferometry processing method as described in claim 1, characterized in that, In step seven, the true phase is obtained after multi-baseline unwrapping. The elevation ambiguity can be used to invert the target elevation data through phase inversion. In the heavy orbit InSAR system, the elevation ambiguity is expressed as: ； It reflects the change with phase. The elevation changes, therefore the target elevation can be calculated using the following formula: ； in, The interference phase is obtained after multi-baseline unwrapping.

Images