A small unmanned aerial vehicle-borne SAR high-precision imaging method for high-slope interferometry

Abstract

The present application belongs to the technical field of synthetic aperture radar, and is a kind of small unmanned aerial vehicle-borne SAR high-precision imaging method for high-slope interferometric measurement. The method is based on the construction of different imaging planes, and the optimal imaging coarse plane and optimal imaging sub-plane are obtained by optimizing the estimation of image indicators, so as to realize high-precision imaging of high-slope, and the subsequent interferometric processing is easy to carry out.

Description

Technical Field

[0001] This invention belongs to the field of synthetic aperture radar technology, specifically, it relates to a small unmanned aerial vehicle (UAV)-borne high-precision SAR imaging method for interferometric measurement of high slopes. Background Technology

[0002] Geological disasters occur frequently worldwide, with landslides being the most dangerous and occurring most frequently among various types. Landslide monitoring and early warning are crucial for disaster prevention and mitigation, and deformation is one of the key characteristics of landslides. Synthetic Aperture Radar (SAR) is a new type of radar with two-dimensional high-resolution imaging capabilities. Combined with phase interferometry, it can measure surface deformation, providing data support for landslide early warning and monitoring. Furthermore, its integration with miniaturized and lightweight unmanned aerial vehicle (UAV) technology makes it more flexible and convenient.

[0003] SAR imaging is fundamental to interferometry. Back projection (BP) is a time-domain imaging algorithm that utilizes the high-precision motion trajectory information of the radar platform and the instantaneous slant range calculated from the spatial coordinates of the imaging area to overlay grid points in the imaging area point by point, achieving high-precision imaging. However, due to the influence of airflow disturbances, the flight paths of small UAVs are complex, making it difficult to fly in a uniform straight line, leading to the failure of azimuth translation invariance. Simultaneously, due to the influence of surface undulations, the spatial elevation of the imaging area is unknown, resulting in errors between the actual slant range history of the target and the antenna phase center (APC) and the imaging slant range history. Furthermore, this slant range error is positively correlated with the target elevation. When the surface elevation is much smaller than the instantaneous slant range, the flat terrain assumption can be used to obtain high-quality imaging results without the aid of an external DEM. However, due to the limited flight altitude of small UAVs, especially when facing high slope scenarios, the terrain to be measured is often close to the UAV's flight altitude. In this case, the surface elevation does not meet the condition of being much smaller than the instantaneous slant range, and the flat terrain assumption no longer holds, leading to problems such as defocusing or distortion in the imaging results, thus introducing a registration challenge. To eliminate slant range errors when using the BP algorithm for imaging processing, an accurate external digital elevation model (DEM) is required. However, obtaining the external DEM of the imaging area is difficult, and its accuracy cannot be guaranteed. Therefore, it is necessary to design a small UAV-borne high-precision SAR imaging method for high slope scenarios, which can also ensure the registration performance of subsequent interferometric measurements. Summary of the Invention

[0004] In view of this, the present invention provides a small UAV-borne SAR high-precision imaging method for interferometric measurement of high slopes, which can achieve accurate focused imaging of areas with elevation.

[0005] To achieve the above-mentioned objectives, the technical solution of this invention is as follows:

[0006] This invention proposes a high-precision SAR imaging method for small unmanned aerial vehicles (UAVs) targeting high slopes, comprising the following steps:

[0007] S1, Construct the imaging plane

[0008] For high slope scenarios, to obtain a plane that best reflects the spatial elevation of the target scene, it is used as the imaging plane during the backpropagation (BP) algorithm imaging process to improve image quality. Imaging planes L with different tilt angles are constructed using the ground distance from the starting point of the imaging region as the axis. n n = 1, 2, ..., N, where N represents the number of imaging planes constructed. The elevation corresponding to the distance x from the ground satisfies the following relationship.

[0009]

[0010] Where x0 represents the distance from the starting point to the imaging plane, and α n The constructed plane L n The tilt angle.

[0011] S2, Plane-by-plane imaging

[0012] The constructed plane L n These are successively used as imaging planes, and the BP algorithm is used to perform imaging processing on these imaging planes in sequence.

[0013] S3, Obtaining the Optimal Coarse Plane

[0014] Image evaluation metrics are used to evaluate the imaging results of different imaging planes, and the imaging plane corresponding to the best imaging result based on the optimal image evaluation metric is taken as the optimal coarse plane. For example, when contrast is used as the image evaluation metric, the method is as follows:

[0015]

[0016] When the image evaluation metric used is minimum entropy, the acquisition method is as follows:

[0017]

[0018] in The optimal coarse planes, C(L), are selected based on contrast and image entropy, respectively. n ), E(L n ) are respectively L n The contrast and image entropy corresponding to the imaging results of a plane.

[0019] S4, Optimal Coarse Plane Partitioning

[0020] Let the obtained optimal coarse plane be L. f When this is initially used as the elevation of the imaging area, the maximum slant distance error resulting from the traditional flat terrain assumption is ultimately expressed as:

[0021]

[0022] Where Δr is the maximum deviation of the UAV from its straight trajectory, and H is the ideal flight altitude of the radar platform. Let x be the elevation value corresponding to the optimal coarse plane at ground distance x, and R be the shortest slant distance from ground distance x on the optimal coarse plane to the ideal track, which can be expressed as:

[0023]

[0024] Where α f This represents the tilt angle of the optimal coarse plane.

[0025] When the maximum residual slope error ΔR max When the resulting phase error is less than π / 4, its impact on imaging is negligible, i.e., ΔR max <λ / 16. Therefore, λ / 16 is used as a threshold to divide the coarse plane into blocks along the distance direction. Now, let's take the scene center distance x. c The corresponding position is the starting point for block division. Sub-blocks are divided along two directions: decreasing ground distance and increasing ground distance. The change in slant distance error in each sub-block area is no greater than λ / 16.

[0026] S5. Optimal Subplane Acquisition

[0027] Using each sub-block defined by the optimal coarse plane as the center, sub-regions with different tilt angles are constructed. For each sub-block divided along the direction of increasing ground distance, sub-planes with different tilt angles are generated with the starting point of ground distance as the axis; for each sub-block divided along the direction of decreasing ground distance, sub-planes with different tilt angles are generated with the ending point of ground distance as the axis. Each sub-plane is imaged one by one, and the imaging plane corresponding to the imaging result with the best image evaluation index is selected as the optimal sub-plane.

[0028] S6, Optimal Subplane Fusion Imaging

[0029] The optimal sub-planes corresponding to all sub-regions are stitched together and fused to obtain a continuous optimal imaging plane. This is then used as the new imaging plane for imaging processing to obtain the optimal imaging result.

[0030] Furthermore, the plane-by-plane imaging method in S2 is a backward projection imaging method.

[0031] Furthermore, the image evaluation metrics in S3 and S5 can be image entropy, contrast, etc.

[0032] Beneficial effects

[0033] This invention provides a small UAV-borne SAR high-precision imaging method for interferometric measurement of high slopes. Based on the construction of different imaging planes, the optimal imaging coarse plane and the optimal imaging sub-plane are obtained by optimizing the estimation of image indicators, thereby realizing high-precision imaging of high slopes and facilitating subsequent interferometric processing. Attached Figure Description

[0034] Figure 1 This is a flowchart illustrating the implementation of the present invention;

[0035] Figure 2 A schematic diagram of constructing the imaging coarse plane;

[0036] Figure 3 The optimal coarse plane temporal imaging model for non-linear flight paths is shown in the figure.

[0037] Figure 4 This is a schematic diagram of subplane processing. Detailed Implementation

[0038] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0039] like Figure 1 As shown, this invention provides a small unmanned aerial vehicle (UAV)-borne high-precision SAR imaging method for interferometric measurement of high slopes, specifically including the following steps:

[0040] S1, Construct the imaging plane

[0041] With the starting point of the imaging region as the axis, construct as follows: Figure 2 The image coarse planes at different tilt angles are shown.

[0042] The nth imaging plane L is constructed for the entire scene. n The tilt angle is α n , there is α n =α1+(n-1)dθ,α1≤α n ≤α N , where α1, α N These are the minimum and maximum values ​​set for the tilt angle of the imaging plane, respectively, where dθ is the step interval angle.

[0043] In the constructed L n On a plane, the elevation corresponding to a distance x from the ground. Satisfying Relationships

[0044]

[0045] Where x0 represents the initial distance from the ground to the imaging plane.

[0046] S2, Plane-by-plane imaging

[0047] The constructed planes are sequentially used as imaging planes, and the BP imaging algorithm is used for imaging processing on these planes. The expression for the imaging result of the scattering point P(x,y,z) in the imaging region is:

[0048]

[0049] in, R n (t m z) is the instantaneous slant range between the scattering point P and the antenna phase center (APC). For scattering point P in coarse plane L n Projection point on Instantaneous slant distance from APC. n (t r ,t m |P) represents the pulse compression result of the echo signal from scattering point P, λ is the carrier wavelength, A is the amplitude after imaging, and t r t m These represent the distance time and the azimuth time, respectively. c The signal center frequency.

[0050] S3, Obtaining the Optimal Coarse Plane

[0051] Image evaluation metrics are used to evaluate the imaging results of different imaging planes, and the imaging plane corresponding to the best imaging result based on the optimal image evaluation metric is taken as the optimal coarse plane. Specific image evaluation metrics include image contrast, image entropy, etc.

[0052] Wherein, contrast ratio is defined as

[0053]

[0054] Where σ(·) represents the variance, E(·) represents the mean, and |I(x,y)| 2 This represents the intensity of the pixel at point (x, y) in the image.

[0055] The greater the image contrast, the better the image focusing effect. Therefore, we calculate the image contrast under different imaging planes and select the image with the highest contrast and its corresponding imaging plane.

[0056] Image entropy is defined as

[0057]

[0058] Where S(X) and I(x,y) represent the image entropy and pixel value, respectively, and N x N yIt represents the total number of pixel units in the distance and orientation directions of the image.

[0059] The lower the image entropy, the more information the image contains and the better its focusing performance; conversely, the higher the entropy, the less information the image contains and the worse its focusing quality. Therefore, we calculate the image entropy under different imaging planes and select the image with the minimum entropy and its corresponding imaging plane.

[0060] S4, Optimal Coarse Plane Partitioning

[0061] In t m The position of APC at time (Δx(t)) m ),vt m ,H+Δz(t m The instantaneous slant range between any scattering point P(x,y,z) within the target region and the APC is...

[0062]

[0063] Wherein, Δx(t) m ), Δz(t) m Let P(x,y,z) represent the motion errors of the radar platform along the X and Z axes, respectively, and R be the shortest slant distance from the scattering point to the ideal trajectory. When the optimal coarse plane is approximated as the elevation value of the target area, the scattering point P(x,y,z) satisfies the following relationship: (10) can be transformed into

[0064]

[0065] The projection point of the scattering point P(x,y,z) onto the horizontal reference plane is P'(x',y',0). Then y' = y, and the instantaneous slant distance between P' and APC is:

[0066]

[0067] Slope error for

[0068]

[0069] Let x - x' = δx represent the position offset in the ground direction, then at the beam center time t c ,have The relationship between ground distance offset and elevation can be obtained as follows:

[0070]

[0071] Substituting (14) into (13), and rearranging, we obtain the expression for the relationship between the slant distance error and the target elevation:

[0072]

[0073] Figure 3 This represents the optimal coarse-plane temporal imaging model under a non-linear trajectory. It assumes the maximum motion error of the radar platform deviating from the flight path is Δr, and the maximum motion errors of the platform along the X and Z axes are Δx = Δrcosθ and Δz = Δrsinθ, where... From the trigonometric function relationship, the maximum slope distance error can be expressed as:

[0074]

[0075] Among them, the imaging plane corresponding to the optimal coarse plane ground distance x satisfies the relationship with the shortest slant distance R of the ideal track. Corresponding elevations satisfy the relationship α f This represents the inclination angle of the optimal coarse plane. From the above, we can see that the maximum slope distance error can be expressed as a function ΔR of the ground distance x. max (x).

[0076] The phase error caused by the maximum slant range error can be ignored when its impact on imaging is less than π / 4, i.e., ΔR must be satisfied. max <λ / 16. Therefore, λ / 16 is used as a threshold to divide the coarse plane into blocks. The distance from the scene center to x... c Starting from the corresponding position, sub-blocks are divided along the directions of decreasing and increasing ground distance, respectively. The change in slant distance error in each sub-block area is no greater than λ / 16.

[0077] A schematic diagram of optimal coarse plane sub-block partitioning and processing is shown below. Figure 4 As shown, the range of ground distances from the i-th sub-region of the scene center. Its satisfaction relationship

[0078]

[0079] Where i can be positive or negative. For the scene ground distance to the right terminal block of the center, i takes a positive value, and for the scene ground distance to the left terminal block of the center, i takes a negative value. The · symbol indicates that the integer part is rounded down to 0.

[0080] It is worth noting that as the ground distance increases, the trend of slope distance error gradually slows down. Therefore, the range of ground distances included in each sub-block gradually increases with the increase of ground distance direction (except for the last sub-block).

[0081] S5. Optimal Subplane Acquisition

[0082] Centered on each sub-block of the optimal coarse plane, sub-planes with different tilt angles are constructed. A schematic diagram of the sub-plane construction is shown below. Figure 4 As shown.

[0083] Starting from the center of the scene, subplane B is drawn along the direction of increasing ground distance. i When constructing a new subplane centered on the starting point, the position corresponding to its distance from the starting point is used as the axis to construct subplanes with different dip angles. The dip angle of the subplane is α. f Centered on α, the range is defined as α. sub ∈[α f -θ,α f +θ], where θ is the maximum tilt angle of the constructed plane relative to the central sub-plane. Therefore, with B i The tilt angle of the nth subplane constructed around the center is α. sub_n =α f -θ+(n-1)δθ, where δθ is the step interval angle.

[0084] Starting from the center of the scene, subplane A is drawn along the direction of decreasing ground distance. i When constructing a new subplane around the center, the position corresponding to the ground distance cutoff point is used as the axis to construct subplanes with different tilt angles. Based on the reference point center as the optimal coarse plane scene center, the tilt angle center of the newly created subplane on the left side of the scene is set to -π+α. f The range of tilt angles is defined as α. sub ∈[-π+α f -θ,-π+α f +θ]. Therefore, with A i The tilt angle of the nth subplane constructed around the center is α. sub_n =-π+α f -θ+(n-1)δθ.

[0085] Each subplane is imaged sequentially, and the imaging plane corresponding to the imaging result with the best image evaluation index is selected as the optimal subplane. Using A... i The optimal subplane obtained with respect to the center is A. if , with B i The optimal subplane obtained with respect to the center is B. if .

[0086] S6, Optimal Subplane Fusion Imaging

[0087] The optimal sub-planes corresponding to all sub-regions are fused according to their positions at the time of block division to obtain continuous optimal imaging planes. Specifically:

[0088] The optimal subplanes are obtained as...,A 2f A 1f B 1f B 2f Since the sub-block distances are continuous, they can be combined in order of distance magnitude to form a new imaging plane, the expression of which is:

[0089] Lop =[...A 2f A 1f B 1f B 2f [,...] (18)

[0090] Using this as the new imaging plane, imaging is performed to obtain the final imaging result.

[0091] In summary, the above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A small unmanned aerial vehicle (UAV)-borne high-precision SAR imaging method for interferometric measurement of high slopes, characterized in that, Includes the following steps: S1. Constructing the imaging plane: Using the starting point along the distance from the ground in the imaging region as the axis, construct imaging planes with different tilt angles. ,in Indicates the number of imaging planes constructed; S2, Plane-by-plane imaging: Imaging the constructed planes These are successively used as imaging planes, and the BP algorithm is used to perform imaging processing on these imaging planes in sequence. S3. Optimal coarse plane acquisition: The imaging results of different imaging planes are evaluated using image evaluation metrics, and the imaging plane corresponding to the imaging result with the best image evaluation metrics is taken as the optimal coarse plane. S4. Coarse plane segmentation: The maximum slope distance error is not less than... As a threshold, the coarse plane is divided into blocks along the distance direction; now, the distance from the scene center is taken as... The corresponding position is the starting point for segmentation. Sub-blocks are formed along both the decreasing and increasing ground distance directions. The variation in slope distance error within each sub-block region is no greater than [missing value]. ; S5. Optimal Subplane Acquisition: Using the sub-blocks divided by each optimal coarse plane as the center, construct sub-regions with different tilt angles; for each sub-block to the right of the scene center, generate sub-planes with different tilt angles with the starting point of ground distance as the axis; for each sub-block to the left of the scene center, generate sub-planes with different tilt angles with the ending point of ground distance as the axis; image each sub-plane one by one, and select the imaging plane corresponding to the imaging result with the best image evaluation index as the optimal sub-plane; S6. Optimal Subplane Fusion Imaging: The optimal subplanes corresponding to all sub-regions are stitched together and fused to obtain a continuous optimal imaging plane. This plane is then used as the new imaging plane for imaging processing to obtain the optimal imaging result.

2. The method for high-precision SAR imaging of small unmanned aerial vehicles for interferometric measurement of high slopes as described in claim 1, characterized in that, The plane-by-plane imaging method in S2 is a back projection imaging method.

3. The method for high-precision SAR imaging of small unmanned aerial vehicles for interferometric measurement of high slopes as described in claim 1, characterized in that, The image evaluation metrics in S3 and S5 are image entropy and contrast.

Images