A passive remote sensing product assisted SAR image mosaic method
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- RES INST OF FOREST RESOURCE INFORMATION TECHN CHINESE ACADEMY OF FORESTRY
- Filing Date
- 2026-05-12
- Publication Date
- 2026-08-04
AI Technical Summary
[0005]本发明要解决的技术问题是针对以上不足，提供一种被动遥感产品辅助的SAR影像镶嵌方法，对现有SAR影像镶嵌技术无法消除不同时间获取的SAR影像间由土壤水分和植被等变化引起的辐射差异的问题，提出了用被动光学植被指数产品和被动微波土壤水分产品计算相邻SAR影像因观测日期不同产生的SAR强度的变化，校准不同观测时间下地表变化带来的辐射异质性，实现不同日期获取的SAR影像辐射异质性的校准技术，最后生成辐射匀质没有色调差异的SAR镶嵌影像
本发明技术方案借助被动微波土壤水分产品和光学植被指数产品计算两个观测日期SAR影像强度的变化量，将不同日期获取的SAR强度校准到相同参考日期，消除了不同获取日期间SAR强度的辐射异质性，再次基础上得到了辐射匀质的SAR镶嵌影像，使得基于本发明得到的镶嵌SAR影像可应用至制图展示、定性以及定量遥感反演等研究中。
Smart Images
Figure CN122510140A_ABST
Abstract
Description
Technical Field
[0001] This invention is a SAR image mosaicking method assisted by passive remote sensing products, belonging to the field of digital geographic information. Background Technology
[0002] High-resolution, large-scale spaceborne synthetic aperture radar (SAR) provides crucial data for monitoring large-scale terrestrial ecosystems and environmental changes. However, spaceborne SAR observations are constrained by satellite revisit cycles, and large-area observation datasets typically consist of multiple images acquired on different dates. These SAR intensity images, acquired at inconsistent times, exhibit significant radiometric differences due to variations in environmental factors such as soil moisture and vegetation. Direct mosaicking would result in significant tonal differences caused by radiometric inhomogeneities, introducing substantial uncertainty in practical applications. To address these issues, it is necessary to calibrate and then mosaic the SAR intensity images acquired from multiple dates.
[0003] Currently, large-scale SAR mosaicking mainly employs methods such as histogram equalization and optimized seam optimization to mitigate radiometric inconsistencies between SAR images. However, when mosaicking SAR images acquired over multiple dates over a large area, this method cannot eliminate radiometric differences between images acquired at different times, resulting in significant radiometric inhomogeneity in the mosaicked images. To address the characteristics of SAR images, a method for generating optimal seams based on image segmentation to extract homogeneous regions is proposed to mitigate radiometric heterogeneity. By calculating the distance weights from overlapping pixels to the effective boundary of their respective image regions, a weighted average of intensity and distance can also weaken mosaicking artifacts. Another approach involves extracting the intensity differences of overlapping pixels in adjacent image orbits and calculating correction coefficients. These correction coefficients are then fitted in both the azimuth and range directions to obtain two-dimensional correction coefficients, which are used to correct SAR intensity images and mitigate radiometric inconsistencies between images. Finally, instead of directly mosaicking SAR intensity images, mosaicking is performed based on statistical parameters (mean, maximum, minimum, and standard deviation) of time-series SAR intensity images. The large-area SAR mosaic images generated by the above methods mainly focus on weakening the radiometric inconsistency of SAR images from different dates by optimizing the stitching strategy. They ignore the real physical differences in SAR intensity caused by changes in environmental factors such as soil moisture and vegetation conditions on different dates, and fail to completely eliminate the radiometric heterogeneity of SAR images acquired on different dates.
[0004] Specifically, the large-area SAR backscattering coefficient mosaic method proposed by Masanobu Shimada and Takahiro Otaki of the Japan Aerospace Exploration Agency (JAXA) has the following basic framework / idea: Figure 1As shown, input two SAR intensity images (first SAR intensity image and second SAR intensity image) after orthorectification and slope compensation. The first step involves setting up multiple equally spaced sampling windows along the azimuth direction in the overlapping area of the two images, and calculating the average intensity of all pixels within the sampling window range of the two images; the second step... The correction coefficient C corresponding to each window is calculated, and multiple correction coefficients can be obtained in the azimuth direction. By fitting multiple correction coefficients in the azimuth direction, continuous correction coefficients in the azimuth direction can be obtained. In the third step, the method is based on the imaging principle of SAR and assumes that the correction coefficient changes linearly with distance in the range direction. Two-dimensional correction coefficients can be obtained for each pixel. Finally, the first SAR intensity image and the second SAR intensity image are corrected by the two-dimensional correction coefficients. Finally, the two images are mosaicked by conventional mosaicking method. This method directly determines the correction scheme by the difference in image intensity, ignoring the radiation heterogeneity caused by changes in environmental parameters such as soil moisture and vegetation. It fails to eliminate radiation differences from the mechanism, resulting in radiation inhomogeneity or radiation distortion problems in the mosaicking results, which is difficult to meet the application requirements of mapping and display, qualitative and quantitative remote sensing parameter inversion. Summary of the Invention
[0005] The technical problem this invention aims to solve is to address the above-mentioned shortcomings by providing a passive remote sensing product-assisted SAR image mosaicking method. This method addresses the issue that existing SAR image mosaicking techniques cannot eliminate radiation differences between SAR images acquired at different times caused by changes in soil moisture and vegetation. It proposes using passive optical vegetation index products and passive microwave soil moisture products to calculate the SAR intensity changes of adjacent SAR images due to different observation dates, calibrating the radiation heterogeneity caused by surface changes at different observation times. This achieves a calibration technique for the radiation heterogeneity of SAR images acquired on different dates, ultimately generating a SAR mosaic image with homogeneous radiation and no tonal differences.
[0006] To solve the above technical problems, the present invention adopts the following technical solution: A passive remote sensing product-assisted SAR image mosaicking method includes the following steps: Step 1, Data Input: Input two SAR intensity images acquired on different dates that have been geocoded and topographically radiometrically corrected, the passive optical vegetation index product corresponding to the observation date, and the passive microwave soil moisture product. The two SAR intensity images are the first SAR intensity image and the second SAR intensity image, respectively. Step 2, determine the spatial overlap area of the images: calculate the spatial range intersection of the two SAR intensity images, and use the spatial range intersection as a mask to extract the SAR intensity of the overlap area; Step 3, extract passive remote sensing data in the overlapping area: Based on the mask, extract passive microwave soil moisture data and optical vegetation index data of the overlapping area corresponding to the observation dates of the two SAR images; Step 4, Construct a radiation correction model: Conduct sensitivity analysis based on the TorVergata model, establish a linear relationship model between SAR intensity change, soil moisture change, and vegetation index, and solve for the model coefficients; Step 5, Full-Image Range Radiometric Calibration: Apply the linear relationship model to the full spatial range of the SAR image to be calibrated, calculate the intensity change of the SAR image to be calibrated relative to the reference date, and complete the radiometric calibration; Step 6, Regional SAR Image Mosaic: The two radiometrically calibrated SAR images are fused using conventional mosaicking methods to output a large-area SAR mosaic image with homogeneous radiometric properties.
[0007] Furthermore, in step 2, the two SAR intensity images are respectively designated as the first SAR intensity image. Second SAR intensity image ; For the first SAR intensity images of the orbital overlap area acquired on two different dates Second SAR intensity image To illustrate, the effective coverage spatial ranges corresponding to the first and second SAR intensity images are read separately. Then, the intersection of the two spatial ranges is calculated as the spatial overlap area of the two images. Finally, using this overlap area as a mask, the intensity of the two images within the mask area is extracted, where the intensity of the mask area extracted from the first SAR intensity image is denoted as... The intensity of the masked area extracted from the second SAR intensity image is denoted as ; This represents the change in SAR intensity in the overlapping area of the images acquired on Date1 and Date2: .
[0008] Furthermore, in step 3, the observation dates of the first SAR intensity image and the second SAR intensity image are Date1 and Date2, respectively. Using Date1 as the reference date, the second SAR intensity image is calibrated to the SAR image of Date1.
[0009] Furthermore, in step 3, the passive microwave soil moisture data obtained from Date1 and Date2 in the masked area are extracted and denoted as... and Changes in soil moisture in the overlapping area of two SAR images It can be represented as: .
[0010] Furthermore, in step 4, the factors dominating SAR intensity changes under surface heterogeneity conditions are simulated and analyzed using the TorVergata model. Surface heterogeneity is expressed by randomly generating a combination of N data points for soil texture, roughness, and vegetation cover. Soil moisture changes are expressed by randomly generating two sets of N soil moisture data points for each set and calculating the difference between the two sets of data. ; Based on the TorVergata model, assuming the surface heterogeneity parameters remain constant, simulating two sets of soil moisture data yields two sets of SAR intensity data. To simulate and obtain the difference between the two sets of SAR intensities.
[0011] Furthermore, the sensitivity analysis in step 4 revealed... and There is a linear relationship between them, and vegetation cover significantly affects and The slope of the linear relationship between them, while the heterogeneity of soil texture and roughness affects... and The influence of the relationship is negligible, and the SAR intensity change is negligible at this time. This can be expressed as: , where P1, P2, P3, and P4 are the coefficients of the linear equation.
[0012] Furthermore, the sensitivity analysis based on the mechanism model will be used to obtain... Applying the calculation formula to the observation data, the change in radar intensity between the first SAR intensity image acquired on Date1 and the second SAR intensity image acquired on Date2 in the orbital overlap area is calculated. Changes in soil moisture and vegetation index The following relationship exists between them: ; P1, P2, P3, P4 are based on and Combined with NDVI data of the orbital overlap area Obtain.
[0013] Furthermore, in step 5, the acquisition Changes in passive microwave soil moisture products over two dates within the geographic area of the image Data and NDVI data, calculation Changes in observation intensity relative to Date1 within the geographic area of the image: ; at this time The unknown intensity on the target date, Date1, is calculated using the following formula: .
[0014] Furthermore, the passive optical vegetation index product is an NDVI product.
[0015] Furthermore, the SAR intensity image is a spaceborne synthetic aperture radar intensity image.
[0016] The present invention adopts the above technical solution and has the following technical effects compared with the prior art: The technical solution of this invention uses passive microwave soil moisture products and optical vegetation index products to calculate the change in SAR image intensity between two observation dates, calibrating the SAR intensity acquired on different dates to the same reference date, eliminating the radiometric heterogeneity of SAR intensity during different acquisition dates, and obtaining a radiometrically homogeneous SAR mosaic image. This allows the mosaic SAR image obtained based on this invention to be applied to research such as mapping, qualitative and quantitative remote sensing inversion. Attached Figure Description
[0017] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.
[0018] Figure 1 This is a schematic diagram of the existing SAR backscattering coefficient mosaic technique in the background art of this invention; Figure 2 This is a schematic diagram of the structure of the large-area SAR image mosaicking method in an embodiment of the present invention. Detailed Implementation
[0019] Examples, such as Figure 2 As shown, a passive remote sensing product-assisted SAR image mosaicking method includes... Step 1, Data Input; The data input includes two SAR intensity images acquired on different dates, a passive optical vegetation index, and a passive microwave soil moisture product corresponding to the date the SAR images were acquired. The two SAR intensity images are the first SAR intensity image and the second SAR intensity image. The SAR intensity images input here have been geocoded and topographically radiometrically corrected.
[0020] Step 2: Determine the spatial overlap area of the images; For the first SAR intensity images of the orbital overlap area acquired on two different dates Second SAR intensity image To illustrate, the effective coverage spatial ranges corresponding to the first and second SAR intensity images are read separately. Then, the intersection of the two spatial ranges is calculated as the spatial overlap area of the two images. Finally, using this overlap area as a mask, the intensity of the two images within the mask area is extracted, where the intensity of the mask area extracted from the first SAR intensity image is denoted as... The intensity of the masked area extracted from the second SAR intensity image is denoted as .
[0021] This represents the change in SAR intensity in the overlapping area of the images acquired on Date1 and Date2: .
[0022] Step 3: Extract passive remote sensing data from the overlapping area; The date Date1, when the first SAR intensity image was acquired, is set as the target date to which SAR images acquired on other dates should be calibrated; the second SAR intensity image is the SAR image acquired on Date2 that needs to be corrected to Date1.
[0023] Based on the mask area obtained in the previous step, passive microwave soil moisture data acquired on Date1 and Date2 of the mask area are extracted respectively, and denoted as... and Changes in soil moisture in the overlapping area of two SAR images It can be represented as: .
[0024] Step 4: Construct a radiation correction model; The factors dominating SAR intensity variation under surface heterogeneity were simulated and analyzed using the Tor Vergata model. This was achieved by randomly generating N data points for soil texture, roughness, and vegetation cover (NDVI). sim The combination of these methods expresses surface heterogeneity by randomly generating two sets of N soil moisture data points each and calculating the difference between the two sets of data to represent changes in soil moisture. ; Based on the Tor Vergata model, assuming the surface heterogeneity parameters remain constant, simulating two sets of soil moisture data yields two sets of SAR intensity data. To simulate and obtain the difference between the two sets of SAR intensities; Sensitivity analysis found and There is a linear relationship between them, and vegetation cover has a significant impact. and The slope of the linear relationship between them, while the heterogeneity of soil texture and roughness affects... and The influence of the relationship is negligible, and the SAR intensity change is negligible at this time. This can be expressed as: , where P1, P2, P3, and P4 are the coefficients of the linear equation.
[0025] Applying the above formula, derived from sensitivity analysis based on the mechanistic model, to the observation data, the change in radar intensity between the first SAR intensity image acquired on Date1 and the second SAR intensity image acquired on Date2 in the orbital overlap region is... Changes in soil moisture and vegetation index The following relationship exists between them: ; P1, P2, P3, P4 are based on and Combined with NDVI data of the orbital overlap area Obtain.
[0026] Step 5, full-image-range radiometric calibration; Get Changes in passive microwave soil moisture products over two dates within the geographic area of the image Using data and NDVI data, calculate the change in intensity relative to Date1 observations within the geographic area of the image: ; at this time The unknown intensity on the target date, Date1, is calculated using the following formula: This eliminates the radiometric differences between the first and second SAR intensity images caused by changes in soil moisture and vegetation conditions at different acquisition times (Date1 and Date2), thus achieving calibration of the radiometric heterogeneity of SAR images at different time phases.
[0027] Step 6: Regional SAR image mosaicking; The calibrated multi-temporal SAR images are mosaicked using conventional mosaicking methods. The first SAR intensity image and the calibrated second SAR intensity image are imported into software such as ArcGIS or ENVI. The overlapping area fusion method is set (such as mean fusion), and feathering smoothing is performed. Finally, a large-area SAR mosaic image is output.
[0028] The description of this invention is given for illustrative and descriptive purposes only and is not intended to be exhaustive or to limit the invention to the forms disclosed. Many modifications and variations will be apparent to those skilled in the art. The embodiments were chosen and described in order to better illustrate the principles and practical application of the invention and to enable those skilled in the art to understand the invention and design various embodiments with various modifications suitable for a particular purpose.
Claims
1. A SAR image mosaicking method assisted by passive remote sensing products, characterized in that: Includes the following steps: Step 1, Data Input: Input two SAR intensity images acquired on different dates that have been geocoded and topographically radiometrically corrected, the passive optical vegetation index product corresponding to the observation date, and the passive microwave soil moisture product. The two SAR intensity images are the first SAR intensity image and the second SAR intensity image, respectively. Step 2, determine the spatial overlap area of the images: calculate the spatial range intersection of the two SAR intensity images, and use the spatial range intersection as a mask to extract the SAR intensity of the overlap area; Step 3, extract passive remote sensing data in the overlapping area: Based on the mask, extract passive microwave soil moisture data and optical vegetation index data of the overlapping area corresponding to the observation dates of the two SAR images; Step 4, Construct a radiation correction model: Conduct sensitivity analysis based on the TorVergata model, establish a linear relationship model between SAR intensity change, soil moisture change, and vegetation index, and solve for the model coefficients; Step 5, Full-Image Range Radiometric Calibration: Apply the linear relationship model to the full spatial range of the SAR image to be calibrated, calculate the intensity change of the SAR image to be calibrated relative to the reference date, and complete the radiometric calibration; Step 6, Regional SAR Image Mosaic: The two radiometrically calibrated SAR images are fused using conventional mosaicking methods to output a large-area SAR mosaic image with homogeneous radiometric properties.
2. The SAR image mosaicking method assisted by passive remote sensing products as described in claim 1, characterized in that: In step 2, the two SAR intensity images are respectively designated as the first SAR intensity image. Second SAR intensity image ; For the first SAR intensity images of the orbital overlap area acquired on two different dates Second SAR intensity image To illustrate, the effective coverage spatial ranges corresponding to the first and second SAR intensity images are read separately. Then, the intersection of the two spatial ranges is calculated as the spatial overlap area of the two images. Finally, using this overlap area as a mask, the intensity of the two images within the mask area is extracted, where the intensity of the mask area extracted from the first SAR intensity image is denoted as... The intensity of the masked area extracted from the second SAR intensity image is denoted as ; This represents the change in SAR intensity in the overlapping area of the images acquired on Date1 and Date2: .
3. The SAR image mosaicking method assisted by passive remote sensing products as described in claim 2, characterized in that: In step 3, the observation dates of the first SAR intensity image and the second SAR intensity image are Date1 and Date2, respectively. Using Date1 as the reference date, the second SAR intensity image is calibrated to the SAR image of Date1.
4. The SAR image mosaicking method assisted by passive remote sensing products as described in claim 2, characterized in that: In step 3, the passive microwave soil moisture data obtained from Date1 and Date2 in the masked area are extracted and denoted as follows: and Changes in soil moisture in the overlapping area of two SAR images It can be represented as: 。 5. The SAR image mosaicking method assisted by passive remote sensing products as described in claim 2, characterized in that: In step 4, the TorVergata model is used to simulate and analyze the factors that dominate SAR intensity changes under surface heterogeneity. Surface heterogeneity is expressed by randomly generating a combination of N soil texture, roughness, and vegetation cover data. Soil moisture changes are expressed by randomly generating two sets of N soil moisture data for each set and calculating the difference between the two sets of data. ; Based on the TorVergata model, assuming the surface heterogeneity parameters remain constant, simulating two sets of soil moisture data yields two sets of SAR intensity data. To simulate and obtain the difference between the two sets of SAR intensities.
6. The SAR image mosaicking method assisted by passive remote sensing products as described in claim 5, characterized in that: Sensitivity analysis in step 4 revealed and There is a linear relationship between them, and vegetation cover has a significant impact. and The slope of the linear relationship between them, while the heterogeneity of soil texture and roughness affects... and The influence of the relationship is negligible, and the SAR intensity change is negligible at this time. This can be expressed as: , where P1, P2, P3, and P4 are the coefficients of the linear equation.
7. The SAR image mosaicking method assisted by passive remote sensing products as described in claim 6, characterized in that: Sensitivity analysis based on mechanism model Applying the calculation formula to the observation data, the change in radar intensity between the first SAR intensity image acquired on Date1 and the second SAR intensity image acquired on Date2 in the orbital overlap area is calculated. Changes in soil moisture and vegetation index The following relationship exists between them: ； P1, P2, P3, P4 are based on and Combined with NDVI data of the orbital overlap area Obtain.
8. The SAR image mosaicking method assisted by passive remote sensing products as described in claim 2, characterized in that: In step 5, the acquisition Changes in passive microwave soil moisture products over two dates within the geographic area of the image Data and NDVI data, calculation Changes in observation intensity relative to Date1 within the geographic area of the image: ； at this time The unknown intensity on the target date, Date1, is calculated using the following formula: .
9. The SAR image mosaicking method assisted by passive remote sensing products as described in claim 1, characterized in that: The passive optical vegetation index product is an NDVI product.
10. The SAR image mosaicking method assisted by passive remote sensing products as described in claim 1, characterized in that: The SAR intensity image is a spaceborne synthetic aperture radar intensity image.