Multi-view sar autonomous positioning method based on joint estimation of range doppler

Abstract

The multi-view SAR autonomous positioning method based on the joint distance Doppler estimation comprises the following steps: unifying multi-view SAR image imaging parameters to the same coordinate system, obtaining the coordinates of a target in each image through an image matching method; constructing a multi-view distance equation set by using the position vector of the synthetic aperture center of each image, the target image coordinates and the distance, and constructing a multi-view Doppler equation set by using the velocity vector of the synthetic aperture center of each image and the target image coordinates; introducing a ranging deviation and a velocity vector deviation as unknown quantities to be solved, correcting the multi-view Doppler equation set, and positioning the target S by using the spatial observation degrees of freedom provided by the multi-view SAR image sequence and considering the ranging and velocity deviation and the velocity deviation.

Description

Technical Field

[0001] This invention belongs to the technical field of airborne synthetic aperture radar, and particularly relates to a multi-view SAR autonomous localization method based on range-Doppler joint estimation. Background Technology

[0002] Synthetic Aperture Radar (SAR) is an active microwave remote sensing device with all-weather, high-resolution imaging and target localization capabilities for regions of interest. SAR autonomous target localization refers to determining the spatial three-dimensional coordinates or latitude and longitude of a target without prior information such as a target reference map or ground control points, using onboard inertial navigation data, radar parameters, and the target's position in the SAR image. This achieves absolute target localization without reference points, enabling intelligence reconnaissance, surveillance, and long-range target targeting and fire support. It boasts advantages such as high timeliness, high flexibility, and universality.

[0003] For target localization in single-scene SAR images, the more mature algorithms include the F. Leberl algorithm, the collinearity equation method, the polynomial method, and the range-Doppler algorithm. Because airborne platforms are more susceptible to disturbances caused by high-altitude airflow, resulting in poor platform stability and difficulty in control, current airborne SAR autonomous localization technology is mainly based on the range-Doppler model proposed by JPL Laboratory, and often uses Earth model equations or incorporates digital elevation models as supporting conditions. Other image localization methods require a certain number of control points to calculate correction parameters.

[0004] Due to the combined effects of radar system errors, inertial navigation errors, atmospheric conditions, and differences between the reference ellipsoid and the actual terrain, as well as varying degrees of geometric distortion such as displacement, stretching, and rotation in SAR images, the autonomous positioning error of current airborne single-scene SAR is relatively large, and the positioning accuracy cannot meet the requirements of coordinate-based attacks. Furthermore, single-scene SAR imaging is susceptible to occlusion and overlapping, making it impossible to acquire image information of the target for positioning.

[0005] To further improve the accuracy of SAR autonomous positioning without relying on external databases, the redundancy of space observation data is utilized based on the spatial geometric relationships of multiple SAR images to enhance positioning accuracy. Multi-view SAR positioning constructs a stereo SAR image of the same scene from multiple different perspectives, combines image matching techniques to obtain target image points that meet the conditions, and then achieves ground target positioning through methods such as orientation parameter calculation under the imaging model and stereo intersection. In 1996, Dowman used ERS-1's precise orbital parameters to calculate orientation parameters, achieving a planar accuracy of four times the pixel resolution for stereo SAR image positioning without ground control points. In 2011, Capaldo first conducted a three-dimensional information extraction experiment using COSMO-skyMed's focused stereo imagery. In 2012, Toutin performed three-dimensional information extraction from Radarsat 2 stereo imagery without ground control points, significantly reducing the need for control points. Although the above-mentioned spaceborne stereo positioning methods achieve high positioning accuracy, they typically give little consideration to the influence of orbital parameter errors and are not highly adaptable to airborne mobile platforms.

[0006] In summary, existing SAR image localization technologies have the following main drawbacks:

[0007] 1) The autonomous positioning accuracy of single-scene SAR is relatively low. When there are no control points or external databases, the autonomous positioning accuracy is easily affected by factors such as parameter calculation accuracy, support condition performance, shadow occlusion and the applicability of the imaging model.

[0008] 2) Multi-view SAR autonomous positioning only uses radar ranging capability to construct a range equation set, without utilizing SAR's Doppler frequency measurement capability.

[0009] 3) Under the condition of ranging and velocity measurement errors, it is difficult to obtain a high-precision least squares solution by only constructing a set of nonlinear positioning equations about the target position.

[0010] In summary, the shortcomings of existing technologies have led to a reduction in radar performance.

[0011] In view of this, the present invention is hereby proposed. Summary of the Invention

[0012] The purpose of this invention is to provide a multi-view SAR autonomous localization method based on range-Doppler joint estimation, addressing the technical problem of reduced radar performance caused by deficiencies in existing technologies. The technical solution of this invention offers numerous beneficial effects, as described below:

[0013] A multi-view SAR autonomous localization method based on joint range-Doppler estimation is provided, the method comprising:

[0014] The imaging parameters of multi-view SAR images are unified into the same coordinate system, and the coordinates of the target in each image are obtained through image matching methods.

[0015] The position vector P of the aperture center is synthesized from each image. i (B i ,L i H i ), target image coordinates s Ii and distance r s,meas,i A multi-view distance equation system was constructed, and the velocity vector V at the aperture center was synthesized using images. i Target image coordinates sI i Constructing a multi-view Doppler equation system;

[0016] Introducing ranging bias r bias and velocity vector deviation V bias As unknowns to be solved, the multi-view Doppler equations are modified.

[0017] The modified multi-view Doppler equations are solved using Newton iteration to output the three-dimensional positioning results.

[0018] Compared with the prior art, the technical solution provided by the present invention has the following beneficial effects:

[0019] To address the issues of low positioning accuracy in single-view SAR and susceptibility to inertial navigation system (INS) accuracy, this invention increases the number of observation equations by leveraging the spatial observation degrees of freedom provided by multi-view SAR image sequences. It constructs a range-Doppler equation system using SAR's ranging and frequency measurement capabilities, borrows the pseudorange concept from satellite positioning, treats constant ranging deviation as an estimator, and incorporates the constant velocity deviation of the INS into the positioning equations. This enables joint estimation of the target's three-dimensional position and ranging / velocity deviations, effectively improving positioning accuracy. This invention does not rely on external support conditions such as Earth ellipsoid models and digital elevation maps. It fully utilizes the range-Doppler non-coplanar capability provided by the multi-view observation configuration and considers the impact of system ranging and INS velocity deviations. It effectively solves the technical challenge of low positioning accuracy in single-view SAR and has significant application potential in intelligence reconnaissance, terrain mapping, and long-range precision strikes. Furthermore, it uses the spatial observation degrees of freedom provided by multi-view SAR image sequences, considering ranging / velocity deviations and velocity deviations, to locate target S. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart of the method of the present invention;

[0022] Figure 2 This is a schematic diagram of multi-view SAR imaging positioning.

[0023] Figure 3 This is a schematic diagram of a multi-view SAR positioning simulation. Detailed Implementation

[0024] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0025] The multi-view SAR autonomous localization method based on joint range-Doppler estimation provided by this invention is described in [reference needed]. Figure 1 The methods include:

[0026] S101: Unify the imaging parameters of multi-view SAR images to the same coordinate system, and obtain the target's coordinates in each image through image matching methods. Specifically:

[0027] Processing of single-view SAR images: When acquiring single-view SAR images from airborne radar, an imaging coordinate system is established with the synthetic aperture center of the SAR image as the origin. The target is the point to be located, and the target's position is calculated using coordinate transformation from the image system to the imaging system, the geographic system, and the geocentric system to obtain its latitude and longitude.

[0028] Due to the constantly changing observation position of the aircraft, the imaging coordinate system in multi-view SAR imaging is not uniform. Therefore, it is necessary to first unify the parameters of each image into the same coordinate system. There are several methods for unifying the coordinate system, such as establishing a unified geographic system based on the scene center.

[0029] Let the positions of the centers of each aperture be (B) i ,L i H i ), where B i It's latitude, L i It's longitude, H. iLet the height be the ellipsoid height. Set the position of the imaging center to (B). c ,L c H c The imaging center can be selected as the center of the scene illuminated by the beam, used to establish a unified imaging coordinate system; only a rough approximation is needed. The velocity of the aperture center is (V... E,i V N,i V S,i The subscript E represents the eastward velocity, N represents the northward velocity, and S represents the skyward velocity.

[0030] With the imaging center (B) c ,L c H c Establish a rectangular coordinate system with the origin as the origin, including the X-axis, y-axis, and z-axis, where the X-axis points east, the Y-axis points north, and the Z-axis points upward. Then, the position vector of each aperture center in this coordinate system is P. i =[P xi ,P yi ,P zi ] T The velocity vector is V i =[V xi V yi V zi ] T T is the transpose, and the transformation relationship is:

[0031]

[0032] Where [ΔX] ic ΔY ic ΔZ ic ] T It is the coordinate of the vector from the aircraft to the imaging center in the ECEF (geocentric-ground-fixed coordinate system) coordinate system.

[0033]

[0034] It is the radius of the circle, R. e It is the semi-major axis of the ellipsoid, e 2 =2f-f 2 Let f be the first eccentricity, and f be the Earth's oblateness.

[0035] The aircraft velocity vector V in the imaging coordinate system i =[V xi V yi V zi ] T With Northeast Sky Speed ​​V ENU The conversion relationship is

[0036]

[0037] Among them, X ecef Y ecef和 Y ecef These are the coordinates in the geocentric-ground-fixed coordinate system.

[0038] 1) in,

[0039]

[0040]

[0041] dz ecef / dL i =0

[0042] 2)

[0043] 3)

[0044]

[0045]

[0046]

[0047]

[0048]

[0049] All multi-view images are imaged using the aforementioned unified coordinate system. The purpose is to match identical scenes in each image and align them. Because SAR images have a depth of focus, the imaging plane is the same, and the height is zero. The target image is the projection of the true position s onto the imaging plane. Ii They have the same distance and Doppler, which is why SAR images exhibit overlapping. Matching corresponding points in an image can be achieved using image registration algorithms such as SIFT, which will not be elaborated upon here.

[0050] S102: Synthesize the position vector P of the aperture center using each image. i (B i ,L i H i ), target image coordinates s Ii and distance r s,meas,i A multi-view distance equation system was constructed, and the velocity vector V at the aperture center was synthesized using images. i Target image coordinates s Ii Constructing a multi-view Doppler equation system, specifically:

[0051] Based on the actual position s of the radar at the center of the aperture and the image position s of the target Ii Since the distances are the same, we can obtain the following system of distance equations:

[0052] Because radar has ranging errors.

[0053] P1 represents the center position of the first image, i and M are natural integers, PM is the Mth sub-image, Si represents the position of the point to be located, and SiM is the position of the point to be located in the Mth image.

[0054] Based on the target's actual position s and image position s Ii (There are M image locations, one image) Figure 1 ( ) have the same Doppler frequency f d

[0055] λ is the wavelength;

[0056] Because |P i -s Ii |=|P i -s|, the above expression can be simplified to V i T s = V i T s Ii Using image coordinates s Ii To avoid directly calculating the Doppler effect, the Doppler equations are established, and the system of Doppler equations is as follows:

[0057]

[0058] S103: Introduce ranging bias r bias and velocity vector deviation V bias As the unknowns to be solved, the multi-view Doppler equations are modified, specifically:

[0059] The distance measurement deviation r cannot be calculated from a single image. bias and velocity vector deviation V bias How to calculate the error is as follows:

[0060] Due to atmospheric transmission errors, system delay errors, and other factors, the distance from the radar to the target measured by SAR is not the true distance. The common ranging error in SAR images from different viewpoints is defined as the ranging deviation r. bias As an unknown quantity to be solved, and due to the limited accuracy of the inertial navigation system, it is assumed that there is a constant velocity offset V of the inertial navigation system during the multi-view SAR imaging time. bias Similarly, this variable is introduced as a variable to be solved into the distance-Doppler equations in step two. The method of this invention considers the existence of errors and makes corrections as follows:

[0061] Let the aperture center times of different images be τ. i Then the ranging deviation r is introduced. bias and speed deviation V bias The corrected distance-Doppler positioning equations are given. By solving the Doppler positioning equations, the ranging error r can be calculated. bias and speed deviation V bias The true location s of the target.

[0062]

[0063] S104: Correct the multi-view Doppler equations and perform Newton iteration to solve the 3D localization results. Specifically:

[0064] The equations in S103 are solved using Newton's iteration method, with a termination condition set for the number of iterations. For example, to solve for the three-dimensional location of the target, the solution process is as follows:

[0065] Rewrite the system of equations in step three in terms of the independent variable. The equation is: FR1 is the constructed distance function, and FD is the Doppler function (1...M).

[0066]

[0067] To achieve iterative solution of the system of equations, initial values ​​are set. The mean of the image coordinates For, r bias0 =0, V bias0 =

[000] T Take the partial derivative with respect to X, and write it as a correction number. A system of linear equations with variables:

[0068] AΔX = b, where: b is the set representation of the FR and FD functions; A is the set representation of the derivatives of FR with three unknowns;

[0069] in

[0070]

[0071]

[0072] and Substituting the function value with initial value X0, the component of the partial derivative with respect to the target position s is:

[0073]

[0074]

[0075]

[0076]

[0077]

[0078]

[0079]

[0080]

[0081] For ranging deviation r bias The components of the partial derivative are:

[0082]

[0083]

[0084] Where θ s ψ is the oblique angle of the ground plane. i Wipe the corners of the floor.

[0085] Regarding the speed measurement deviation V bias The components of the partial derivative are:

[0086]

[0087]

[0088]

[0089]

[0090]

[0091]

[0092]

[0093]

[0094] Then, the least squares method is used to solve for the correction number ΔX. Ls =(A T A) -1 A T b.

[0095] Determine ΔX Ls Is it less than a specified threshold (e.g., Δs≤10)? -2 r bias ≤1m, |Δv bias |≤10 -3If the value is greater than the threshold, the initial value is corrected: X0 = X0 + ΔX Ls The process continues until the maximum number of iterations is exceeded; if the number of iterations is less than the threshold, the target position in the initial value X0 is converted into latitude, longitude, and altitude and then output as the final three-dimensional positioning result of the target.

[0096] Figure 2 This is a geometrical schematic diagram of multi-view SAR imaging and positioning according to the present invention, wherein the carrier aircraft observes the same scene from M perspectives, obtaining M SAR images. The synthetic aperture center position vector of each image is P. i The velocity vector is V i Let i = 1, 2, ..., M. The large circle at point A in the image represents the true location s of the target. After image sequence matching, the coordinates of the same target in different images are si. Ii The point is represented by a small dot at position B.

[0097] Figure 3 This is a schematic diagram of a multi-view SAR positioning simulation. Asterisks indicate the aircraft's position. There are a total of M=7 observation views, arranged in a spiral curve. Triangles represent the target position with coordinates (300, 200, 100). The imaging plane height is zero. A ranging bias r is set. bias =20m, speed measurement deviation V bias = [0.1 -0.2 0] m / s, aircraft position vector P i Velocity vector V i and image position s Ii See Table 1.

[0098]

[0099] Table 1 Simulation parameters for multi-view SAR positioning

[0100] The target location estimated using this method is: The estimated distance measurement deviation is Speed ​​measurement deviation estimate is Compared to the previous method, the root mean square error of planar localization of scattering points in each image was reduced from 79.46m to 7.04m.

[0101] The product provided by this invention has been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this invention. The descriptions of the embodiments above are merely for the purpose of helping to understand the core ideas of this invention. It should be noted that those skilled in the art can make various improvements and modifications to the invention without departing from the principles of the invention, and these improvements and modifications also fall within the protection scope of the invention claims.

Claims

1. A multi-view SAR autonomous localization method based on range-Doppler joint estimation, characterized in that, The methods include: The imaging parameters of multi-view SAR images are unified into the same coordinate system, and the coordinates of the target in each image are obtained through image matching methods. Synthesize the position vector of the aperture center using each image. Target image coordinates and distance A multi-view distance equation system was constructed, and the velocity vector at the aperture center was synthesized using images. Target image coordinates Constructing a multi-view Doppler equation system; Introducing distance measurement bias and velocity vector deviation The distance-Doppler equations are modified as unknowns to be solved. The modified multi-view Doppler equations are solved using Newton iteration to output the three-dimensional positioning results.

2. The multi-view SAR autonomous localization method based on range-Doppler joint estimation according to claim 1, characterized in that, The imaging parameters of multi-view SAR images are unified into the same coordinate system, and the coordinates of the target in each image are obtained through image matching methods, including: The processing of a single-scene SAR image includes: when acquiring a single-scene SAR image from an airborne radar, establishing an imaging coordinate system with the synthetic aperture center of the SAR image as the origin; Unify all image parameters to the same coordinate system, including: Let the positions of the centers of each aperture be respectively ,in It's latitude. It's longitude. Let be the height of the ellipsoid, then the location of the radar's imaging center is... The velocity at the center of the aperture is expressed as In this context, the subscript E represents the eastward velocity, N represents the northward velocity, and S represents the skyward velocity. With the center of the aperture Establish a rectangular coordinate system with the origin as the origin. The X-axis of the rectangular coordinate system points eastward, the Y-axis points northward, and the Z-axis points upward. The position vector of each aperture center in the rectangular coordinate system is: The velocity vector is T is the transpose, and the position vector transformation relationship is expressed as: in, It is the coordinate of the vector from the carrier aircraft to the imaging center in the Earth-centered Earth-fixed coordinate system (ECEF coordinate system); The velocity vector transformation relationship is expressed as: in, It is the radius of the circle. It is the semi-major axis of the ellipsoid. For the first eccentricity, It is the Earth's oblateness; Aircraft velocity vector in imaging coordinate system With Northeast Speed The transformation relationship is expressed as: in, In the formula, X ecef， Y ecef and Y ecef The coordinates are in the Earth-centered Earth-fixed coordinate system, Nc is the radius of the prime meridian at the imaging center, and n is the coordinates. east_c n north_c and n up_c These are the unit vectors from the Earth's center to the imaging center in the East, North, and Sky directions, respectively, in the ECEF coordinate system. The subscript c represents the imaging center. Multi-view images are imaged in a unified coordinate system.

3. The multi-view SAR autonomous localization method based on range-Doppler joint estimation according to claim 2, characterized in that, Synthesize the position vector of the aperture center using each image. Target image coordinates and distance A multi-view distance equation system was constructed, and the velocity vector at the aperture center was synthesized using images. Target image coordinates Constructing a multi-view Doppler equation system, including: Based on the actual position of the radar at the center of the aperture to reach the target and image location Since the distances are the same, the system of distance equations can be obtained, which includes: The synthetic aperture center position vector for each image is: i and M are natural integers, PM is the Mth subgraph, Si represents the position of the point to be located, and S IM The location of the point to be located in the Mth image; Based on the target's true location and image location Having the same Doppler frequency f d , We can obtain, Wavelength; Furthermore, Simplify to and using image coordinates Establishing the Doppler equations, we derive the following system of Doppler equations: 。 4. The multi-view SAR autonomous localization method based on range-Doppler joint estimation according to claim 3, characterized in that, Introducing distance measurement bias and velocity vector deviation As unknowns to be solved, the modified multi-view Doppler equations include: The common ranging error in SAR images from different perspectives is defined as ranging bias. and constant velocity offset As unknowns to be solved, the variables to be solved are introduced into the distance-Doppler equations in step two; Let the aperture center times of different images be respectively This introduces a ranging bias. and speed deviation The corrected range-Doppler positioning equations are: 。 5. The multi-view SAR autonomous localization method based on range-Doppler joint estimation according to claim 4, characterized in that, The modified multi-view Doppler equations are solved using Newton iteration to output three-dimensional localization results, including: The modified range-Doppler positioning equations are transformed into equations relating the independent variable... The equation, , of which F R1 For the constructed distance function, F D The Doppler function; To achieve iterative solution of the system of equations, initial values ​​are set. The mean of the image coordinates for, , ,right Taking the partial derivative, we get... With correction number A system of linear equations with variables; because, Where: b is F R and F D Set representation of functions; A is F R Set representation for differentiating three unknowns; and Substitute initial values The function value for the target position Find the components of the partial derivative; For distance measurement deviation The components of the partial derivative are: ,in, This is an oblique angle view of the ground plane. Wipe the corners of the floor; Speed ​​measurement deviation Find the components of the partial derivative; Released, revised The least squares method is used to solve for... ; judge Is it less than the specified threshold? If so, change the initial value. After the target location is converted into latitude, longitude, and altitude, it is output as the final 3D positioning result of the target. Otherwise, the initial value is used. Make corrections, the modification method is as follows Then proceed to the next iteration, until the maximum number of iterations is exceeded, and then the corrected initial value is returned. The target location is converted into latitude, longitude, and altitude, and then output as the final three-dimensional positioning result of the target.

Images