Forest vertical structure information inversion method and system based on radar interferometry

Abstract

The invention provides a forest vertical structure information inversion method and system based on radar interferometry, and the method comprises the steps: carrying out the small-view processing of high-resolution SAR interferometry data with a space baseline, calculating the small-view InSAR phase height, extracting the statistical distribution of the small-view interferometry phase height in a set space window, and carrying out the calculation of the small-view interferometry phase height. Generating a small-looks interferometric phase height histogram; establishing an under-forest ground elevation regression estimation model based on the statistical characteristics of the small-looks interferometric phase height histogram, and determining model parameters to invert the terrain; and two canopy top indexes are extracted, two index results are fused by setting a threshold value, and final forest vertical structure information is generated. The method has the advantages that more canopy internal and under-forest information can be captured only by using single-baseline and single-polarization radar interference data, and the sensitivity to a forest vertical structure is improved; and the forest vertical section, the tree height and the terrain can be synchronously inverted.

Description

Technical Field

[0001] The present application belongs to the field of forest remote sensing mapping, and specifically relates to a method and system for inverting forest vertical structure information based on radar interferometry. The method is a technical method for inverting forest vertical structure information (including vertical structure profile, forest height and underlying surface terrain elevation) by utilizing high-resolution (e.g., satellite-borne L-band dual-station single-polarization single-baseline) interferometric synthetic aperture radar (InSAR) data with a spatial baseline, combined with satellite-borne lidar data calibration. Background Art

[0002] As a vital component of terrestrial ecosystems, medium- to high-resolution (<100 m) measurements of forest biophysical parameters, such as tree height and biomass, are crucial for monitoring tropical forest heterogeneity, estimating carbon stocks, and assessing ecosystem function. Spaceborne remote sensing is the primary means of achieving these goals. Commonly used observation technologies include optical sensors, lidar, and synthetic aperture radar (SAR) systems.

[0003] Spaceborne lidar (such as NASA's GEDI and ICESat-2 / ATLAS) can provide high-precision observations of tree height at discrete points, but their sampling mode is discrete, making it difficult to achieve continuous mapping over large areas. Optical remote sensing sensors (such as the Landsat series) rely on the reflectivity of vegetation and have insufficient penetration into dense forests, making it impossible to obtain structural information below the canopy. The bistatic interferometry measurement technology in synthetic aperture radar systems has unique advantages in inverting the three-dimensional structure of forests due to its microwave penetration, precise altimetry, and sensitivity to vertical structure. For example, Germany's TanDEM-X satellite uses a bistatic interferometry mode, but the system uses the X-band (~3.1 cm), which has limited penetration into forests. Moreover, most of the data uses a single-baseline, single-polarization observation mode, making the polarization interferometry or tomography methods, which are the mainstream international methods for measuring forest tree height, understory topography, and vertical structure profiles, unsuitable.

[0004] China's Lutan-1 satellite mission is the world's first spaceborne L-band dual-station Interferometric Synthetic Aperture Radar (InSAR) mission. Its primary data acquisition mode, similar to the TanDEM-X mission, remains single-polarization, single-baseline, providing a new data source for detecting forest vertical structure. The L-band (15-30cm wavelength) has greater forest penetration, capturing more information within the canopy and understory. However, methods for inferring forest vertical structure based on this single-polarization, single-baseline dual-station InSAR data are currently lacking. Existing methods often rely on physical scattering model assumptions and are poorly adaptable to single-polarization, single-baseline data. Therefore, it is urgent to develop adaptive technical solutions to fully utilize the advantages of data acquired by Lutan-1 and satellites with similar observation regimes. Summary of the Invention

[0005] The purpose of this application is to overcome the shortcomings of the existing technology in terms of continuous mapping, single-baseline single-polarization data adaptability and inversion accuracy.

[0006] To achieve the above objectives, this application proposes a method for inverting forest vertical structure information based on radar interferometry, including:

[0007] Step S1: Perform low-look-number processing on high-resolution (e.g., L-band dual-station single-polarization single-baseline) SAR interferometry data with a spatial baseline, calculate the low-look-number InSAR phase height using the phase-height conversion formula, extract the statistical distribution of the low-look-number interferometry phase height within a set spatial window, and generate a low-look-number interferometry phase height histogram;

[0008] Step S2: Based on the statistical characteristics of the small-view interferometer phase height histogram, a regression estimation model for the ground elevation under the forest is established. A differentiation strategy is used to determine the model parameters for different laser sample scenes to invert the terrain.

[0009] Step S3: Combine the low-view interferometry phase height histogram with the lidar waveform analysis method to extract two canopy top indicators in the histogram. Then, set a threshold to fuse the two indicator results to generate the final forest vertical structure information.

[0010] As an improvement to the above method, the small number of views is 1-4 views.

[0011] As an improvement to the above method, the phase-height conversion formula is:

[0012] h φ =Δφ / κ z

[0013] Among them, h φis the InSAR phase height with small look-number; Δφ is the interferometric phase difference, κ z is the interference vertical wave number.

[0014] As an improvement to the above method, the set spatial window is 20-50m.

[0015] As an improvement to the above method, the understory ground elevation regression estimation model is:

[0016] h g =α·μ-β·σ

[0017] Among them, h g is the ground elevation under the forest; α and β are regression coefficients, i.e., model parameters; μ and σ are the mean and standard deviation of the small-view interferometer phase height histogram, respectively.

[0018] As an improvement to the above method, the method of determining model parameters using a differentiated strategy for different laser sample scenarios includes:

[0019] Set α = 1, allowing β to vary with space;

[0020] When the number of laser sampling points corresponding to a single-view radar image is greater than the set sampling point number threshold, the statistical relationship between the coefficient β and the tree height is established using the lidar measurement of the tree height, and the coefficient β is determined through a linear regression model;

[0021] When the laser sampling points corresponding to a single-view radar image are less than or equal to the set sampling point number threshold, the statistical relationship between the coefficient β and σ is established using the small-view interferometry phase height standard deviation σ as a feature.

[0022] As an improvement to the above method, step S2 further includes: optimizing the inverted terrain in combination with the lowest peak of the understory terrain in the histogram, including:

[0023] For bare land and medium-high vegetation, if there is a minimum peak in the small view count histogram within the range of the terrain inversion value plus or minus the set range, it is considered that the minimum peak corresponds to the optimized terrain inversion value;

[0024] For dwarf vegetation, if the histogram still has a value but does not have a minimum peak in the interval of the terrain inversion value plus or minus the set range, the position where the second-order derivative of the histogram in this interval is 0 is taken as the optimized terrain inversion value;

[0025] For super-high vegetation, if the histogram has no value within the range of the terrain inversion value plus or minus the set range, the terrain inversion value is directly used as the optimized terrain inversion value.

[0026] As an improvement to the above method, the two canopy top indicators include: RH98 and a canopy top indicator of a 3dB attenuation point.

[0027] The present application also provides a forest vertical structure information inversion system based on radar interferometry, which is implemented based on the above method and includes:

[0028] The module generates a small-look-number interferometric phase height histogram. It is used to perform small-look-number processing on high-resolution (e.g., L-band dual-station single-polarization single-baseline) SAR interferometric measurement data with a spatial baseline. The module calculates the small-look-number InSAR phase height using the phase-height conversion formula, extracts the statistical distribution of the small-look-number interferometric phase height within a set spatial window, and generates a small-look-number interferometric phase height histogram.

[0029] The terrain inversion module is used to establish a regression estimation model for the ground elevation under the forest based on the statistical characteristics of the small-view interferometer phase height histogram. It uses a differentiated strategy to determine the model parameters for different laser sample scenes and invert the terrain.

[0030] A forest vertical structure information generation module is used to combine the small-view interferometer phase height histogram and the lidar waveform analysis method to extract two canopy top indicators in the histogram. The two indicator results are fused by setting a threshold to generate the final forest vertical structure information.

[0031] Compared with the prior art, the advantages of this application are:

[0032] 1. This application is based on high-resolution SAR interferometry data with a spatial baseline (specifically, spaceborne L-band dual-station single-polarization single-baseline). The L-band has strong forest penetration, can capture more information within the canopy, and improves sensitivity to forest vertical structure.

[0033] 2. The proposed "small-view interferometry phase height histogram" method does not rely on physical scattering model assumptions and is highly adaptable to single-polarization and single-baseline data, reducing the need for complex parameters.

[0034] 3. It can achieve large-scale continuous mapping. Based on data acquired by the domestically produced Lutan-1 satellite (the world's first spaceborne L-band dual-station interferometric radar satellite) or other satellites with similar observation systems, it can generate forest vertical structure information products, breaking through the limitations of discrete sampling of spaceborne lidar.

[0035] 4. High inversion accuracy. For tropical forests up to 45m high, the tree height inversion accuracy is about 5m (relative error 10-15%). The terrain inversion error varies with slope, ranging from 3m (<2°), 4m (2°-6°), 7m (6°-25°), and 9m (>25°), meeting China's DEM accuracy standards.

[0036] 5. It can simultaneously invert forest vertical profiles, tree heights, and terrain, providing multi-dimensional data support for forest biomass estimation, ecological change monitoring, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 Shown is a flow chart of the forest vertical structure information inversion method based on radar interferometry;

[0038] Figure 2 Shown is a schematic diagram of the underlying surface topography (DTM) inversion strategy;

[0039] Figure 3 Shown is a schematic diagram of the fusion of forest height inversion indicators

[0040] Figure 4 Shown are histograms of bare land and forest generated using real data from Lutan-1;

[0041] Figure 5 The figure shows a schematic diagram of terrain inversion optimization using real data from Lutan-1, which distinguishes four histogram types: bare land, short vegetation, medium-tall vegetation, and very tall vegetation. DETAILED DESCRIPTION

[0042] The technical solution of this application is described in detail below with reference to the accompanying drawings.

[0043] This application aims to utilize high-resolution (such as spaceborne L-band dual-station single-polarization, single-baseline) SAR interferometry data with a spatial baseline, combined with spaceborne lidar data for calibration, to achieve high-precision, large-scale inversion of forest vertical structure information (including vertical structure profiles, forest height, and underlying surface terrain elevation) through the "small-view interferometry phase height histogram" method.

[0044] The method of this application is not only applicable to spaceborne L-band dual-station radar interferometry data, such as China's LuTan-1. As long as it is an advanced radar interferometry satellite, whether it is a dual-station (one transmitter, two receivers) or a heavy-orbit (single transmitter, single receiver) interferometry radar satellite, as long as there is a certain spatial baseline between the two observations of the interferometry measurement, so as to be sensitive to the vertical structure information of the forest, and at the same time have high resolution (the high resolution mentioned in this application is a resolution of less than 6m) to support radar waves penetrating forest gaps (forest windows), the method of this application can be applied.

[0045] Example 1

[0046] like Figure 1 As shown, the present application proposes a method for inverting forest vertical structure information based on radar interferometry, which includes:

[0047] Step 1: Generate a small-look-number interferometric phase height histogram: Perform small-look-number (1-4-look) processing on the high-resolution (less than 6 m) dual-station L-band InSAR single-look complex (SLC) data to balance noise suppression and spatial heterogeneity preservation; calculate the small-look-number InSAR phase height using the phase-height conversion formula, which is:

[0048] h φ =Δφ / κ z (1)

[0049] Where Δφ is the interference phase difference, κ z is the interferometric vertical wavenumber. By combining the height of ambiguity (HoA) to correct for periodic jumps and extracting the statistical distribution of the small apparent interferometric phase height within a larger spatial window (20-50m), a histogram is generated to represent the vertical structure profile of the forest. This histogram can be compared with the waveform of a full-waveform lidar (such as GEDI) to verify its ability to capture vertical stratification features.

[0050] Step 2: Invert the underlying surface topography (DTM): Figure 2 As shown in Figure 2, based on the statistical characteristics (mean μ, standard deviation σ) of the small-view interferometer phase height histogram, a regression estimation model for the ground elevation under the forest is established:

[0051] h g =α·μ-β·σ (2)

[0052] Among them, h gis the ground elevation under the forest, and (α, β) are the regression coefficients. In the following, we assume that α = 1 and allow β to vary spatially. A differentiated strategy is adopted for different laser sample scenarios. In scenarios with sufficient laser samples (a single radar image can contain tens of thousands of laser sampling points), a statistical relationship (positive correlation) between coefficient β and tree height is established, using the lidar tree height as a feature. This coefficient β is determined using a linear regression model (bare ground, β = 0; tall tree, β ≈ 2). In scenarios with limited laser samples (a single radar image contains only a few hundred to a few thousand laser sampling points), a statistical relationship (positive correlation) between coefficient β and σ is established, using the standard deviation σ of the interferometric phase height with a small number of appearances. Initial values ​​for terrain inversion are obtained, and combined with the lowest peak value of the understory terrain in the histogram, four histogram types (bare ground, short vegetation, medium-tall vegetation, and very tall vegetation) are distinguished to optimize terrain inversion. The details are as follows: for bare land and medium and high vegetation, in the interval near the initial value of terrain inversion (plus or minus a small amount), if there is a minimum peak in the small viewpoint histogram, then the minimum peak (first-order derivative is 0) is considered to correspond to the optimized terrain inversion value; for short vegetation, in the interval near the initial value of terrain inversion, the histogram still has a value but does not have a minimum peak, then the position where the second-order derivative of the histogram in this interval is 0 is taken as the optimized terrain inversion value; for super-high vegetation, in the interval near the initial value of terrain inversion, if the histogram has no value, the initial value is directly used instead of the optimized terrain inversion value.

[0053] Step 3: Invert forest height: Figure 3 As shown in the figure, based on the small-view interferometer phase height histogram and drawing on the lidar waveform analysis method, two canopy top indicators are extracted, namely RH98 (98% relative height, referring to the lidar RH98 definition) and 3dB attenuation point (the power attenuation point of the first peak at the top of the canopy in the reference histogram); the two indicator results are fused through the threshold, for example, 7.5m is used as the threshold, the 3dB attenuation point result is used below this value, and the RH98 result is used above this value to improve the inversion accuracy.

[0054] The proposed method for inverting forest vertical structure information based on radar interferometry is compatible with data from the domestic LuTan-1 satellite and can be extended to other similar high-resolution InSAR data with spatial baselines. The following uses LuTan-1 satellite data as an example to verify the proposed method:

[0055] Step 1: Generate a small-view interferometer phase height histogram

[0056] The Lutan-1 single-look complex (SLC) data are processed with a small view (e.g., 2x2, i.e., 4 views) to generate an interferogram. Within a spatial window (e.g., 50 m), the phase is converted to height using formula (1) and a phase-height histogram is generated. The histogram segment size (which can be set to 0.7-1 m) needs to balance the vertical profile details and noise. The generated histogram can be compared with the waveform of a space-borne lidar (e.g., GEDI), indicating that it can effectively capture the vertical stratification characteristics of forests.

[0057] The histograms of bare land and forest generated using the real data of Lutan-1 are as follows: Figure 4 The histogram of bare land has a more concentrated narrow Gaussian distribution, while the histogram of forest vegetation shows more obvious asymmetric and hierarchical structure characteristics.

[0058] Step 2: Invert the underlying surface terrain elevation (DTM)

[0059] Based on the statistical characteristics (mean μ, standard deviation σ) of the small-view interferometry phase height histogram, a regression estimation model for the ground elevation under the forest is established, as shown in formula (2).

[0060] Differentiated processing is performed based on the number of spaceborne lidar samples in the study area. For areas with sufficient laser samples, a statistical regression model of coefficient β and tree height is established (bare ground β = 0, tall tree β ≈ 2), using the tree height interpolation results of the spaceborne laser to determine the spatial distribution of coefficient β. This is substituted into the model formula (2) to calculate the DTM. For areas with sparse laser samples, a statistical regression model of coefficient β and σ is established using the standard deviation σ of the low-level interferometric phase height as a characteristic to perform an initial estimate. The terrain inversion is optimized by combining the lowest peak of the understory terrain in the histogram to distinguish four histogram types: bare ground, short vegetation, medium-tall vegetation, and very tall vegetation.

[0061] like Figure 5 As shown in Figure 1, the specific distinction between the four histogram types is as follows: for bare land (Type I: σ is small, generally around 2-3 meters) and medium-to-high vegetation (Type III: σ is large), in the interval near the initial value of the terrain inversion (plus or minus a small amount, see the red dashed line segment in the figure), if the small view count histogram has a minimum peak (corresponding to the red dot in the figure) above the noise threshold (corresponding to the green dashed line in the figure), then this minimum peak (first-order derivative is 0) is considered to correspond to the optimized terrain inversion value; for short vegetation (Type II: σ is slightly smaller, slightly larger than Type I), in the interval near the initial value of the terrain inversion, the histogram still has values ​​above the noise threshold but does not have a minimum peak, then the position where the second-order derivative of the histogram in this interval is 0 is taken as the optimized terrain inversion value; for very high vegetation (Type IV: σ is large, similar to Type III), in the interval near the initial value of the terrain inversion, if the histogram has no value above the noise threshold, then the initial value is directly used instead of the optimized terrain inversion value.

[0062] Step 3: Invert forest height

[0063] Based on the small-view interferometer phase height histogram and drawing on the lidar waveform analysis method, two canopy top indicators are extracted from the histogram, namely RH98 (98% relative height, referring to the lidar RH98 definition) and 3dB attenuation point (the power attenuation point of the first peak at the top of the canopy in the reference histogram); the two indicator results are fused through a threshold, for example, using 7.5m as the threshold, the 3dB attenuation point result is used below this value, and the RH98 result is used above this value to improve the inversion accuracy.

[0064] Example 2

[0065] The present application also provides a forest vertical structure information inversion system based on radar interferometry, which is implemented based on the above method and includes:

[0066] The module generates a small-look-number interferometric phase height histogram. It is used to perform small-look-number processing on high-resolution (e.g., spaceborne L-band dual-station single-polarization single-baseline) SAR interferometric measurement data with a spatial baseline. The module calculates the small-look-number InSAR phase height using the phase-height conversion formula, extracts the statistical distribution of the small-look-number interferometric phase height within a set spatial window, and generates a small-look-number interferometric phase height histogram.

[0067] The terrain inversion module is used to establish a regression estimation model for the ground elevation under the forest based on the statistical characteristics of the small-view interferometer phase height histogram. It uses a differentiated strategy to determine the model parameters for different laser sample scenes and invert the terrain.

[0068] A forest vertical structure information generation module is used to combine the small-view interferometer phase height histogram and the lidar waveform analysis method to extract two canopy top indicators in the histogram. The two indicator results are fused by setting a threshold to generate the final forest vertical structure information.

[0069] The present application may also provide a computer device comprising: at least one processor, memory, at least one network interface, and a user interface. The various components in the device are coupled together via a bus system. It will be understood that the bus system is used to enable communication between these components. In addition to a data bus, the bus system also includes a power bus, a control bus, and a status signal bus.

[0070] The user interface may include a display, a keyboard, or a pointing device, such as a mouse, a trackball, a touchpad, or a touch screen.

[0071] It can be appreciated that the memory in the embodiments disclosed in the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (Read-Only Memory, ROM), a programmable read-only memory (Programmable ROM, PROM), an erasable programmable read-only memory (Erasable PROM, EPROM), an electrically erasable programmable read-only memory (Electrically EPROM, EEPROM) or a flash memory. The volatile memory can be a random access memory (Random Access Memory, RAM) used as an external cache. By way of example but not limitation, many forms of RAM are available, such as static random access memory (Static RAM, SRAM), dynamic random access memory (Dynamic RAM, DRAM), synchronous dynamic random access memory (Synchronous DRAM, SDRAM), double data rate synchronous dynamic random access memory (Double Data Rate SDRAM, DDR SDRAM), enhanced synchronous dynamic random access memory (Enhanced SDRAM, ESDRAM), synchronous link dynamic random access memory (Synchlink DRAM, SLDRAM) and direct memory bus random access memory (Direct Rambus RAM, DRRAM). The memory described herein is intended to include but not limited to these and any other suitable types of memory.

[0072] In some embodiments, the memory stores elements, executable modules or data structures, or a subset thereof, or an extended set thereof: an operating system and an application program.

[0073] Among them, the operating system includes various system programs, such as a framework layer, a core library layer, a driver layer, etc., for implementing various basic services and processing hardware-based tasks. The application program includes various application programs, such as a media player (Media Player), a browser (Browser), etc., for implementing various application services. The program for implementing the method of the embodiments of the present disclosure can be included in the application program.

[0074] In the above-mentioned embodiments, the processor can also be used to process the program or instruction stored in the memory, in particular, the program or instruction stored in the application program, by calling the program or instruction.

[0075] Perform the steps of the above method.

[0076] The above method can be applied to or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by hardware integrated logic circuits in the processor or by software instructions. The above processor may be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The above-disclosed methods, steps, and logic block diagrams can be implemented or executed. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the above-disclosed method can be directly implemented and executed by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. The software module can be located in a storage medium well-known in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, etc. The storage medium is located in the memory, and the processor reads the information in the memory and, in conjunction with its hardware, completes the steps of the above method.

[0077] It is understood that the embodiments described herein may be implemented using hardware, software, firmware, middleware, microcode, or a combination thereof. For hardware implementation, the processing unit may be implemented in one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers, microprocessors, or other electronic units or combinations thereof for performing the functions described herein.

[0078] For software implementation, the technology of the present application can be implemented by executing the functional modules (e.g., procedures, functions, etc.) of the present application. The software code can be stored in a memory and executed by a processor. The memory can be implemented in the processor or external to the processor.

[0079] The present application may also provide a non-volatile storage medium for storing a computer program. When the computer program is executed by a processor, each step in the above method embodiment can be implemented.

[0080] Finally, it should be noted that the above embodiments are intended only to illustrate the technical solutions of this application and are not intended to limit the scope of the present invention. Although this application has been described in detail with reference to the embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions to the technical solutions of this application do not depart from the spirit and scope of the technical solutions of this application and should be encompassed by the claims of this application.

Claims

1. A method for inverting forest vertical structure information based on radar interferometry, comprising: Step S1: Perform low-look-number processing on high-resolution SAR interferometry data with a spatial baseline, calculate the low-look-number InSAR phase height using the phase-height conversion formula, extract the statistical distribution of the low-look-number interferometry phase height within a set spatial window, and generate a low-look-number interferometry phase height histogram; Step S2: Based on the statistical characteristics of the small-view interferometer phase height histogram, a regression estimation model for the ground elevation under the forest is established. A differentiation strategy is used to determine the model parameters for different laser sample scenes to invert the terrain. Step S3: Combine the low-view interferometry phase height histogram with the lidar waveform analysis method to extract two canopy top indicators in the histogram. Then, set a threshold to fuse the two indicator results to generate the final forest vertical structure information.

2. The method for inverting forest vertical structure information based on radar interferometry according to claim 1, characterized in that: The small view count is 1-4 view counts.

3. The method for inverting forest vertical structure information based on radar interferometry according to claim 1, characterized in that: The phase-to-height conversion formula is: h φ =Df / k z Among them, h φ is the InSAR phase height with small look-number; Δφ is the interferometric phase difference, κ z is the interference vertical wave number.

4. The method for inverting forest vertical structure information based on radar interferometry according to claim 1, characterized in that: The set space window is 20-50m.

5. The method for inverting forest vertical structure information based on radar interferometry according to claim 1, characterized in that: The regression estimation model of the ground elevation under the forest is: h g =a·m-b·s Among them, h g is the ground elevation under the forest; α and β are regression coefficients, i.e., model parameters; μ and σ are the mean and standard deviation of the small-view interferometer phase height histogram, respectively.

6. The method for inverting forest vertical structure information based on radar interferometry according to claim 1, characterized in that: The method of adopting a differentiated strategy to determine model parameters for different laser sample scenarios includes: Set α = 1, allowing β to vary with space; When the number of laser sampling points corresponding to a single-view radar image is greater than the set sampling point number threshold, the statistical relationship between the coefficient β and the tree height is established using the lidar measurement of the tree height, and the coefficient β is determined through a linear regression model; When the laser sampling points corresponding to a single-view radar image are less than or equal to the set sampling point number threshold, the statistical relationship between the coefficient β and σ is established using the small-view interferometry phase height standard deviation σ as a feature.

7. The method for inverting forest vertical structure information based on radar interferometry according to claim 1, characterized in that: The step S2 further includes: optimizing the inverted terrain in combination with the lowest peak of the understory terrain in the histogram, including: For bare land and medium-high vegetation, if there is a minimum peak in the small view count histogram within the range of the terrain inversion value plus or minus the set range, it is considered that the minimum peak corresponds to the optimized terrain inversion value; For dwarf vegetation, if the histogram still has a value but does not have a minimum peak in the interval of the terrain inversion value plus or minus the set range, the position where the second-order derivative of the histogram in this interval is 0 is taken as the optimized terrain inversion value; For super-high vegetation, if the histogram has no value within the range of the terrain inversion value plus or minus the set range, the terrain inversion value is directly used as the optimized terrain inversion value.

8. The method for inverting forest vertical structure information based on radar interferometry according to claim 1, characterized in that: The two canopy top indicators include: RH98 and a canopy top indicator at a 3dB attenuation point.

9. A forest vertical structure information inversion system based on radar interferometry, implemented based on the method according to any one of claims 1 to 8, characterized in that: The system comprises: The module generates a small-look-number interferometric phase height histogram, which is used to perform small-look-number processing on high-resolution SAR interferometric measurement data with a spatial baseline, calculate the small-look-number InSAR phase height using the phase-height conversion formula, extract the statistical distribution of the small-look-number interferometric phase height within a set spatial window, and generate a small-look-number interferometric phase height histogram. A terrain inversion module is used to establish a regression estimation model for ground elevation under the forest based on the statistical characteristics of the low-view interferometer phase height histogram. It uses a differentiated strategy to determine model parameters for different laser sample scenes and invert the terrain; and A forest vertical structure information generation module is used to combine the small-view interferometer phase height histogram and the lidar waveform analysis method to extract two canopy top indicators in the histogram. The two indicator results are fused by setting a threshold to generate the final forest vertical structure information.

Images