Intelligent perception and early warning system for soil erosion in a basin based on space-air-ground cooperation

Abstract

The present application relates to the technical field of radar remote sensing monitoring and soil and water conservation, and discloses a basin soil erosion intelligent sensing and early warning system based on space-air-ground cooperation, comprising: a synthetic aperture radar observation platform, a plurality of variable-depth active coherent beacons arranged on the surface of a monitored basin, and a data processing and inversion center; the synthetic aperture radar observation platform is used for constructing an air-ground active observation link of downlink radar pulse irradiation and uplink echo reception; the variable-depth active coherent beacon comprises a deep anchoring unit and a surface floating unit, and is used for capturing a surface vertical displacement difference value and mapping the displacement difference value into a baseband phase modulation amount by using a digital gain coefficient greater than 1. By introducing orthogonal coding modulation technology in the slow time domain and combining a matching projection operator based on a precise orbit state vector, accurate decoupling of the beacon echo under a strong ground clutter background is realized.

Description

Technical Field

[0001] This invention relates to the field of radar remote sensing monitoring and soil and water conservation technology, specifically to a watershed soil erosion intelligent sensing and early warning system based on air-space-ground collaboration. Background Technology

[0002] Soil erosion in watersheds is a key factor leading to land degradation, river siltation, and ecological deterioration. Long-term, continuous, and detailed monitoring of soil erosion across large watersheds is crucial for ecological restoration and disaster prevention and mitigation. Traditional soil erosion monitoring relies primarily on manually setting up runoff plots or using stake-insertion methods for fixed-point measurements. While these methods can achieve high single-point accuracy, they are time-consuming and labor-intensive, and the observed data are highly spatially dispersed, making it difficult to reflect the spatial distribution patterns of erosion across the complex topography of the entire watershed.

[0003] With the development of remote sensing technology, optical remote sensing has been widely used for land cover monitoring. However, it is susceptible to cloud and rain weather conditions and struggles to penetrate vegetation canopies to obtain high-precision elevation change information, making it unable to directly quantify the depth of microscopic soil erosion. Synthetic Aperture Radar Interferometry (InSAR) technology, with its all-weather, high-precision land deformation monitoring capabilities, has become a highly promising monitoring method. However, in practical applications of soil erosion monitoring, InSAR technology faces significant challenges. On the one hand, soil erosion areas are often accompanied by drastic changes in vegetation cover or surface dielectric constant, leading to a severe decrease or even complete loss of the temporal coherence of radar echo signals, making it difficult to extract the interferometric phase. On the other hand, soil erosion typically manifests as slow changes at the millimeter or even sub-millimeter level. Limited by atmospheric delay phase and system thermal noise interference, conventional InSAR technology struggles to directly distinguish the minute erosion signals submerged in noise.

[0004] To address the issue of coherence loss, existing technologies have attempted to introduce artificial corner reflectors as coherent targets. While traditional passive corner reflectors can provide stable scattering centers, their radar cross-section is geometrically dependent. To achieve a sufficient signal-to-noise ratio, large-scale portable metal structures are often required, making large-scale deployment in the field difficult. Furthermore, passive reflectors lack signal encoding capabilities, and their echo signals are easily superimposed with strong ground clutter in complex backgrounds, making effective separation and extraction at the signal processing level challenging. More critically, traditional reflectors only reflect mechanical displacement at physical locations and cannot amplify the sensitivity of minute erosion deformations, resulting in insufficient system perception of early-stage, minute erosion. Simultaneously, current monitoring methods largely focus on acquiring single-point data, lacking effective spatial interpolation and field reconstruction mechanisms. This makes it difficult to transform discrete, high-precision monitoring points into a continuous erosion depth distribution map across the entire watershed, limiting the application value of monitoring data in comprehensive watershed management. Therefore, this invention provides a watershed soil erosion intelligent sensing and early warning system based on a space-air-ground collaborative approach to address the shortcomings of existing technologies. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a watershed soil erosion intelligent sensing and early warning system based on air-space-ground collaboration, which solves the problems mentioned in the background technology.

[0006] This invention provides a watershed soil erosion intelligent sensing and early warning system based on air-space-ground collaboration. The system aims to solve the technical problems of existing remote sensing technology being unable to accurately monitor minute surface erosion deformation in vegetated areas, and traditional single-point monitoring being unable to reflect the spatial distribution of erosion across the entire watershed.

[0007] The system mainly consists of a synthetic aperture radar (SAR) observation platform, several variable-depth active coherent beacons deployed on the surface of the monitored watershed, and a data processing and inversion center. The SAR observation platform, acting as the active observation source, is responsible for transmitting pulse signals to the monitoring area and receiving echoes. The variable-depth active coherent beacons capture soil erosion depth using the relative displacement between deep anchored units and surface floating units. This vertical displacement difference is converted into a baseband phase modulation quantity, and an orthogonal coding sequence is loaded in the slow time domain. The signal, carrying orthogonal coding characteristics and physical quantity features, is then forwarded back to the radar platform via an RF amplification link. The data processing and inversion center receives the raw echo data, constructs a projection operator matching the beacon characteristics, decouples the beacon's observation phase from strong background clutter, and then inverts to generate a soil erosion depth distribution map for the entire watershed.

[0008] In terms of specific hardware implementation and signal generation mechanism, this invention employs an active modulation scheme to overcome the limited radar cross-section of passive reflectors. The variable-depth active coherent beacon converts physical displacement into electrical signals through a high-precision displacement sensing unit. To improve the detection sensitivity to minute erosions, the system is configured with a digital gain coefficient, set as a constant greater than 1, which linearly amplifies the phase response in the digital domain, enabling micrometer-level mechanical displacement to be mapped into phase changes that can be significantly identified by radar. Simultaneously, the RF front-end module integrates a low-noise amplifier and a power amplifier, and achieves transmit / receive isolation through an RF circulator or transceiver switch, ensuring that the beacon can capture radar pulses in real time, amplify them with high gain, and perform phase modulation forwarding, thereby establishing a high signal-to-noise ratio signal component in the synthetic aperture radar echo.

[0009] During the signal processing stage, the data processing and inversion center utilizes the characteristics of slow-time domain orthogonal coding for coherent demodulation. For each beacon node, the system constructs a dedicated matching projection operator, which combines the beacon's unique orthogonal coding sequence with the instantaneous geometric slant range calculated based on the radar's precise orbital state vector. To eliminate range cell errors caused by discrete sampling, the system employs an interpolation resampling algorithm on the range-axis pulse compressed data to extract complex signal values ​​that precisely correspond to the target's instantaneous position. Through coherent integration on the slow-time axis, the matched beacon signal gains energy, while surface background clutter lacking coding characteristics is suppressed due to incoherent accumulation effects, thus achieving accurate separation of target signals against complex terrain backgrounds.

[0010] In the physical quantity inversion and spatial reconstruction stage, the system first utilizes reference beacon nodes deployed in stable areas to calculate the atmospheric phase screen through planar or quadratic surface models to correct atmospheric delay errors during electromagnetic wave propagation. To address phase ambiguity caused by long-term monitoring, a time-dimensional phase unwrapping strategy is employed to recover accumulated phase changes. Finally, to obtain continuous erosion distribution across the entire watershed, the system uses a co-kriging interpolation algorithm. The discrete erosion depths calculated from each beacon node are used as the main variable, while surface texture features extracted from synthetic aperture radar amplitude images and slope data calculated from digital elevation models are used as co-variables. By utilizing the spatial correlation between variables, weighted estimations are performed for areas without beacon deployments, generating rasterized spatial distribution data of erosion depth, achieving full-coverage monitoring of soil and water loss in the watershed.

[0011] This invention provides a smart sensing and early warning system for watershed soil erosion based on a space-air-ground collaborative approach. It has the following beneficial effects: 1. This invention introduces orthogonal coding modulation technology in the slow time domain and combines it with a matching projection operator based on a precise orbital state vector to achieve accurate decoupling of beacon echoes against a strong surface clutter background. By utilizing the orthogonality of the coding sequence and the coherent integral gain, it can effectively suppress incoherent background clutter interference, thereby overcoming the defect of traditional synthetic aperture radar interferometry technology that is prone to losing coherence in areas with dense vegetation or drastic changes in surface dielectric properties. This ensures stable acquisition and high signal-to-clutter ratio extraction of soil erosion signals in complex field environments.

[0012] 2. This invention employs an active transponder architecture that combines mechanical displacement sensing with electronic phase modulation. By introducing a digital gain coefficient greater than 1, it amplifies the sensitivity to physical erosion depth. By linearly amplifying the phase response in the digital domain, it breaks the inherent constraint that the phase change in traditional passive radar reflection measurement is limited by the physical displacement. This enables synthetic aperture radar to detect micro-millimeter-level surface erosion changes that are much smaller than its range resolution, significantly improving the system's early perception capability and monitoring accuracy for minor soil erosion processes.

[0013] 3. This invention utilizes the co-kriging interpolation algorithm to achieve spatial fusion and field reconstruction of multi-source geographic information, solving the problem that discrete single-point monitoring cannot reflect the erosion distribution of the entire watershed. The system uses high-precision point erosion data obtained from beacons as the main variable, and radar image texture and digital elevation model slope data with full coverage characteristics as co-variables. It uses the spatial cross-correlation between variables to perform weighted estimation of blank areas without nodes, thereby constructing a continuous and detailed watershed-level soil erosion depth distribution map, providing comprehensive and quantitative data support for regional soil and water conservation management. Attached Figure Description

[0014] Figure 1 This is a schematic diagram illustrating the overall application scenario and architecture of the system according to the present invention; Figure 2 This is a flowchart of the method of the present invention; Figure 3 This is a block diagram of the beacon module structure of the present invention; Figure 4 This is a flowchart of the virtual tomography decoupling process based on orthogonal projection of the present invention. Detailed Implementation

[0015] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] Please see the appendix Figure 1 -Appendix Figure 4 The present invention provides a watershed soil erosion intelligent sensing and early warning system based on air-space-ground collaboration. The system mainly includes: a synthetic aperture radar observation platform, several variable depth active coherent beacons distributed in the watershed, and a data processing and inversion center. Synthetic Aperture Radar (SAR) observation platforms are deployed on satellites or UAVs to transmit linear frequency modulated pulse signals into the monitored area and receive backscattered echoes from the Earth's surface. These SAR observation platforms are configured with a fixed pulse repetition frequency and define an azimuthal slow time axis for signal processing. Variable-depth active coherent beacons are arrayed on the surface of the watershed to be monitored, serving as both the sensing end and the signal modulation end for physical quantities of surface soil erosion. Specifically, the variable-depth active coherent beacon comprises deep-anchored units, surface-floating units, high-precision displacement sensing units, and phase-coded modulation units. The deep anchoring unit is fixed to a stable bedrock layer or non-erosion layer below the surface, providing a spatial reference. The surface floating unit is placed on the soil surface and undergoes vertical displacement as the surface elevation decreases due to soil erosion. The high-precision displacement sensing unit connects the deep anchoring unit and the surface floating unit to measure the relative displacement change between them, which corresponds to the soil erosion depth at the monitoring point. The phase-coded modulation unit interacts electromagnetically with the high-precision displacement sensing unit and the synthetic aperture radar observation platform. The phase-coded modulation unit receives radar waves emitted by the synthetic aperture radar observation platform, maps the soil erosion depth data measured by the high-precision displacement sensing unit into the phase and amplitude characteristics of the echo signal according to a preset slow-time-domain orthogonal coding protocol, and modulates the reflected signal. The data processing and inversion center is communicatively connected to the synthetic aperture radar observation platform to acquire raw radar echo data and perform imaging calculations. The data processing and inversion center includes a virtual tomography decoupling module, a phase extraction and correction module, and an erosion field reconstruction module. The virtual tomography decoupling module performs orthogonal projection integration on the echo data in the slow time domain, separating the modulated signal of the variable-depth active coherent beacon from the surface background clutter. The phase extraction and correction module calculates the absolute phase from the separated signal and removes atmospheric errors. The erosion field reconstruction module uses the calculated discrete-point erosion data to extrapolate the continuous erosion distribution across the entire watershed.

[0017] This invention also provides a method for intelligent sensing and early warning of watershed soil erosion based on air-space-ground collaboration, comprising the following steps: S100, System initialization and air-to-ground timing synchronization; When the synthetic aperture radar observation platform enters the airspace above the monitoring area, the variable depth active coherent beacon enters the working state, the phase coding modulation unit captures the leading edge signal of the radar pulse and locks the pulse repetition frequency, and determines the orthogonal coding sequence corresponding to the current synthetic aperture integration time according to the preset protocol. S200, in-situ sensing and mapping of soil erosion physical quantities; high-precision displacement sensing unit reads the vertical displacement difference between deep anchoring unit and surface floating unit in real time, and phase encoding modulation unit calculates the corresponding baseband phase modulation amount based on the vertical displacement difference. S300, Slow time domain orthogonal coding modulation and signal forwarding; During the scanning duration of the synthetic aperture radar observation platform, the variable depth active coherent beacon uses the corresponding symbols in the orthogonal coding sequence to perform phase or amplitude weighting on the reflected signal for the received radar pulse, and simultaneously superimposes the baseband phase modulation amount calculated in step S200 to form a modulated echo and transmit it to the synthetic aperture radar observation platform. S400, Virtual Tomography and Signal Decoupling; The data processing and inversion center receives the total echo signal composed of the superposition of surface background clutter and modulated echo. The virtual tomography decoupling module constructs a matched filter operator for each variable-depth active coherent beacon, performs weighted integration on the total echo signal on the slow time axis, and uses the orthogonality of the coded sequence to suppress the background clutter signal and extract the beacon signal component. S500, full-basin erosion field reconstruction; the phase extraction and correction module extracts the measurement phase from the beacon signal components and removes the atmospheric delay phase by combining known reference points, and calculates the precise soil erosion depth of each monitoring point. The erosion field reconstruction module uses the soil erosion depth of each monitoring point as the spatial control point and combines the surface texture features of radar images to generate a spatiotemporal distribution map of soil erosion in the entire basin.

[0018] The technical details of each step described above will be explained in detail below with reference to specific embodiments.

[0019] The system initialization and air-to-ground timing synchronization described in step S100 involves signal synchronization and parameter configuration between the variable-depth active coherent beacon and the synthetic aperture radar observation platform. To ensure that the ground node can accurately load the modulated signal within the slow time window of the radar's azimuth scan, the following sub-steps are included: S101. Radar Beam Energy Sensing and System Wake-up. The variable-depth active coherent beacon is in a low-power standby state, with its internal RF front-end envelope detection circuit continuously monitoring the ambient electromagnetic field. When the synthetic aperture radar observation platform scans to the edge of the monitoring area, and the beam signal strength emitted by the radar antenna exceeds the threshold level set by the voltage comparator in the envelope detection circuit, the comparator outputs a high-level trigger signal. The power management module responds to this trigger signal, switching from sleep mode to operating mode to supply power to the digital control unit and phase modulator. S102. Pulse repetition frequency locking and timing alignment. After wake-up, the variable-depth active coherent beacon uses a built-in high-frequency clock counter to capture the rising edges of the received continuous radar pulse sequence. The system records the arrival times of multiple consecutive pulses and obtains the current radar pulse repetition interval through differential operations. This locks the pulse repetition frequency. Lock-based The beacon predicts the theoretical arrival time of the next radar pulse. During this process, the digital control unit reads pre-stored inherent processing delay parameters from the theoretical arrival time, subtracts these inherent delays from the actual arrival time, and generates a modulation enable signal with a preset lead. This modulation enable signal defines an effective time window covering the radar pulse width, ensuring that the phase modulation action and the arrival of the radar pulse are strictly coincident on the time axis. S103. Determination of the slow-time domain orthogonal coding sequence: After locking the pulse timing, the system determines the coding sequence used for the current synthetic aperture integration period. This coding sequence is defined on the slow time axis, that is, for a series of pulses transmitted by the radar at different azimuth positions, the beacon sequentially loads different modulation states. The synthetic aperture radar observation platform is set to integrate the synthetic aperture at a single point on the ground for a given time period. The total number of effective pulses emitted by the radar during that time period is... The product of the synthesis aperture time and the pulse repetition frequency is rounded down. The system selects a length of... The set of discrete orthogonal coded sequences ,in Greater than or equal to For the first A variable-depth active coherent beacon, which, based on a pre-set device ID index or received downlink control commands, retrieves data from a set of... The corresponding encoded sequence vector is called in the middle. The encoded sequence satisfies the discrete orthogonality condition in the slow time domain. In this embodiment, the encoded sequence uses a Walsh-Hadamard sequence, with sequence elements consisting of +1 and -1. The phase-coded modulation unit controls the on / off state of the PIN diodes or RF switches on the RF channel according to the values ​​of the sequence elements, realizing the RF signal in the 0-phase state. Rapid switching between phase states; S104. Spatial Allocation and Multiplexing of Coding Sequences. To avoid signal interference between different beacons within the beam coverage area, the system adopts a geographic grid-based coding allocation strategy. The monitored watershed is divided into several non-overlapping coding multiplexing zones. Within the coverage area sufficiently illuminated by the same radar beam, different variable-depth active coherent beacons are assigned different orthogonal code indices. For two beacons geographically separated by more than one synthetic aperture length, since their Doppler histories do not overlap temporally, the system configures them to reuse the same coding sequence.

[0020] The core of the in-situ sensing and mapping of soil erosion physical quantities described in step S200 lies in establishing a deterministic and quantitative relationship between surface micro-deformation and radar echo electromagnetic characteristics. To achieve accurate conversion from mechanical displacement to electromagnetic phase, the following sub-steps are included: S201. Mechanical-Electrical Signal Conversion of Soil Erosion Depth. A deep-anchored unit of a variable-depth active coherent beacon is fixed to the non-eroding soil layer or bedrock interface, serving as a zero-deformation reference. A surface-floating unit adheres closely to the soil surface, undergoing vertical settlement as soil erosion occurs. A high-precision displacement sensing unit, employing a linear variable differential transformer or magnetostrictive displacement sensor, connects the deep-anchored unit and the surface-floating unit, converting the relative physical position changes of both into analog voltage signals. The system's internal analog-to-digital converter acquires the analog voltage signals at a preset sampling rate and uses a moving average filtering algorithm or digital low-pass filter in the microprocessor to filter out high-frequency vibration components caused by wind and grass movement, extracting the quasi-static displacement value representing the cumulative soil erosion depth. That is, the difference in vertical displacement; S202. Mathematical mapping from physical displacement to baseband phase: To enable the radar to sense this minute displacement via interferometry, the phase-coded modulation unit uses a microprocessor to calculate the corresponding baseband phase modulation amount based on pre-stored synthetic aperture radar carrier wavelength parameters. This baseband phase modulation amount... The calculation follows the following linear model: In the formula, The center carrier wavelength of the synthetic aperture radar transmitted signal; This is the digital gain coefficient (electromechanical coupling coefficient) of the device. In the normal mode, it is set to 1. In the micro-erosion monitoring mode, this coefficient is set to a constant greater than 1, and the phase response sensitivity to displacement is amplified by software algorithm. This is the inherent phase offset constant measured during the equipment's factory calibration. This formula clarifies the specific algorithm logic for calculating the baseband phase modulation amount based on the vertical displacement difference in the claims. S203. Quantization and control word generation of baseband phase. Calculated baseband phase modulation amount. For continuous floating-point values. The phase-coded modulation unit determines the value based on the bit width resolution (e.g., 6-bit or 8-bit) of the internal digital phase shifter or vector modulator. Discretization and quantization are performed to generate the corresponding binary control word. The microprocessor sends this control word to the driver circuit of the RF front end via a digital-to-analog converter (DAC) or a general-purpose input / output (GPIO) interface. This operation introduces a base phase bias in the RF path corresponding to the current erosion depth. This base phase bias remains constant or changes slowly during a single synthetic aperture integration time, serving as the reference phase for subsequent quadrature coding modulation.

[0021] Regarding the slow-time-domain orthogonal coding modulation and signal forwarding described in step S300, this step is a crucial step in realizing air-to-ground cooperative virtual tomography. In this process, the variable-depth active coherent beacon generates an echo signal with specific coding characteristics by amplifying and modulating the amplitude of each radar pulse within the integral time of the synthetic aperture radar. This includes the following sub-steps: S301. Generation and counting of pulse-level modulation timing. During the continuous scanning of the monitoring area by the synthetic aperture radar observation platform, a field-programmable gate array (FPGA) or high-speed microcontroller inside the variable-depth active coherent beacon performs cyclic counting based on a locked pulse repetition frequency. The system maintains a slow-time index variable. This variable corresponds to the radar's first transmission at the current moment. Each pulse, with a value range of [value missing]. to ,in This represents the total number of pulses within the synthesis aperture integration time. Whenever a synchronization trigger signal indicating the arrival of a radar pulse is detected, the digital logic unit initiates a modulation sequence for the current pulse. S302, joint solution of composite modulation phase. The microprocessor calculates based on the current slow-time index. Read the orthogonal encoded sequence allocated to the node from memory. The first in Individual code value Simultaneously, the baseband phase modulation amount representing the soil erosion depth, generated and latched in step S200, is read. The system uses digital addition to fuse the encoded and erosion information to calculate the total modulation phase required for the current pulse. This embodiment uses binary phase-shift keying to load orthogonal codes. If the code elements... To +1, introduce a 0-degree phase shift; if the symbol -1, introduced (i.e., a 180-degree phase shift). The calculation logic for the total modulation phase is as follows: After the calculation is completed, the system will It is converted into the corresponding drive voltage or control level and applied to the phase modulation device of the RF front end; S303. Amplification, Modulation, and Retransmission of RF Signals. The receiving antenna of the variable-depth active coherent beacon captures the radar incident pulse signal, which enters the internal RF link via an RF circulator or transceiver switch. The signal is first amplified by a low-noise amplifier (LNA) to improve the signal-to-noise ratio; then it enters a digitally controlled phase shifter. The digitally controlled phase shifter modulates the overall phase... Under the control of [unclear], a corresponding phase shift is superimposed on the radio frequency signal. The phase-modulated radio frequency signal enters the power amplifier (PA) for power gain processing, so that the signal strength reaches the preset radar cross section level. Finally, the amplified signal is radiated back into space again through the circulator or transmitting antenna. In the slow time domain Above, the first Equivalent radar cross section of each beacon node It is expressed as a time-varying function: In the formula, The total electronic gain of the active link; This refers to the physical aperture cross-section of the antenna; The unit is the imaginary unit. This formula shows that the echo signal carries rapidly changing coded features on the slow time axis. Furthermore, its complex arguments contain the erosion phase information to be measured; S304. Spatial superposition and transmission of hybrid echo signals. The beacon signal, after active modulation and relay, propagates in space and is vector-superimposed with the natural background scattering signals generated by the surrounding surface (soil, vegetation, rocks), ultimately being captured by the receiver of the synthetic aperture radar observation platform. The radar receives the first... The baseband echo signal of each pulse Expressed as: In the formula: For distance to fast time; This is background clutter signal, which occurs in slow time. The above exhibits low-frequency Doppler variation characteristics and does not possess... High-frequency coding transition structure; This represents the total number of beacons within the beam. For the first The amplitude factor of a beacon reaching the radar receiver includes two-way path attenuation and the beacon's own active gain. The envelope of the linear frequency modulated signal transmitted by the radar; For radar and the first Instantaneous slant distance between beacons; This represents the speed of electromagnetic wave propagation.

[0022] Regarding the virtual tomography and signal decoupling described in step S400, this step is performed by a high-performance computing unit deployed in the data processing and inversion center. Its core lies in using a pre-set slow-time-domain orthogonal coding sequence to construct a multi-dimensional matched filter, separating the beacon echo from surface background clutter at the signal processing level, including the following sub-steps: S401. Range-direction pulse compression processing. After acquiring the raw radar echo data, the data processing center performs matched filtering in the fast time domain (range direction). The computing unit retrieves the linear frequency modulated waveform parameters of the radar transmitted signal and constructs a matched filter template. The raw echo data is transformed to the frequency domain using a Fast Fourier Transform (FFT), and a complex multiplication operation is performed with the spectrum of the matched filter template. Then, it is transformed back to the time domain using an Inverse Fast Fourier Transform (IFFT). This processing step compresses the wide pulse echo into a narrow pulse signal, outputting a range-compressed signal. It exhibits a peak response in the form of a Sinc function at fast time points corresponding to the target reflection point; S402, Construction of the target node projection operator. To extract information about a specific beacon from a mixed echo containing multi-source signals, the system performs projection on each known variable-depth active coherent beacon. A dedicated matching projection operator is constructed. The computing unit reads the satellite's precise orbital state vector and obtains the radar platform's state at each slow time step. The three-dimensional spatial coordinates, combined with the beacon geographic coordinates Calculate the instantaneous geometric slant range between the radar platform and the beacon. Projection operator The mathematical expression is constructed as follows: In the formula, For the orthogonal coded sequence assigned to this beacon, the first... The complex conjugate of each code element; The center frequency of the radar carrier; The speed of light. This operator includes a coded sequence term for demodulation and a phase correction term for compensating for the Doppler phase history; S403, Slow-Time Domain Orthogonal Projection Integral. Using the constructed projection operator, the system performs coherent accumulation on the range-compressed signal along the slow-time axis. This is based on the calculated instantaneous geometric slant range. The corresponding echo arrival times are typically located between the discrete sampling points of the compressed signal. The computational unit uses the Sinc interpolation algorithm or the cubic spline interpolation algorithm to extract the discrete distance compressed data. Medium resampling obtains accurate correspondence The complex signal value at time t. Subsequently, the computational unit interpolates the signal value and uses the projection operator. Perform dot product and apply it over all integrals within the synthetic aperture time. The results of each pulse are accumulated and summed. The calculation formula is: In the formula, This indicates an interpolation resampling operation. During this integration process, the target beacon signal achieves in-phase superposition due to precise phase history compensation and coded sequence matching, with an amplitude gain of approximately [missing value]. The background clutter signal, lacking corresponding orthogonal coding characteristics, exhibits incoherent accumulation during the integration and summation process, and its energy is suppressed relative to the beacon signal.

[0023] S404, Extraction of complex decoupling values ​​and enhancement of signal-to-noise ratio. After the above projection integration operation, the calculation unit outputs a value specific to the beacon. complex integral value The argument of this complex value includes the phase of soil erosion being measured. The atmospheric transmission phase and the system's fixed phase path; the amplitude of this complex value reflects the strength of the beacon signal. Through orthogonal coding and matched filtering, the output result... The signal-to-clutter ratio (SCR) is improved compared to the original echo, allowing the system to extract beacon phase information from strong ground clutter interference. This complex value... It is stored in the database as input data for subsequent erosion depth inversion.

[0024] The whole-basin erosion field reconstruction described in step S500 is performed by the data processing and inversion center. Its purpose is to eliminate errors introduced by the propagation medium and to extrapolate the single-point measurements of discrete nodes into a continuous spatial distribution across the entire basin. This includes the following sub-steps: S501. Phase Separation and Atmospheric Error Estimation. The observed phase value is extracted from the complex decoupling value output in step S400. This observed phase consists of the deformation phase caused by soil erosion, the atmospheric delay phase generated by electromagnetic waves crossing the troposphere and ionosphere, and measurement noise. To separate the atmospheric delay phase, the system identifies pre-set reference beacon nodes. These reference nodes are deployed on stable bedrock at the watershed edge, with their surface floating units in a mechanically locked state and zero physical displacement. The computational unit extracts the observed phase of the reference nodes, treating it as a superposition of atmospheric delay and system residual error. The system establishes a planar phase model or a quadratic surface phase model, and uses the least squares method to fit the phase data of the reference nodes, solving for the model coefficients. The fitted Atmospheric Phase Screen (APS) model... Expressed as spatial coordinates Functions: In the formula, The coefficients represent the fitting values. This model quantifies the low-frequency spatial variation trend of atmospheric delay within the monitoring area at the time of radar observation; S502. Correction of net erosion phase and inverse calculation of physical quantities: The calculation unit uses the constructed atmospheric phase screen model to perform point-by-point correction of the observed phase of all non-reference beacon nodes within the monitoring area. The observed phase of each beacon minus the estimated atmospheric phase value calculated by the model at that location. The net erosion phase is obtained. To address the phase ambiguity issue in long-term time series monitoring, the system compares the net erosion phase of the current moment with that of the previous moment. If the difference between the two exceeds [a certain threshold], [the system will proceed accordingly]. If phase entanglement occurs, it can be determined by adding or subtracting... Phase unwrapping is performed in the time dimension at integer multiples to recover the accumulated phase change. Then, based on the phase mapping model determined in step S200, the corrected net erosion phase is obtained through inverse operation. Converted to physical erosion depth The calculation formula is: In the formula, The radar carrier wavelength; The digital gain coefficient preset for the beacon; To calibrate the phase constant of the system.

[0025] S503. Spatial interpolation based on multi-source fusion. To obtain a continuous soil erosion distribution map of the entire watershed, the computational unit uses the discrete erosion depth calculated by each beacon node. As the main variable, surface texture features are extracted from amplitude images generated by synthetic aperture radar, and slope data calculated by a digital elevation model (DEM) are used as covariates. The system employs a co-kriging interpolation algorithm to first calculate the experimental semivariogram of the main variables and the cross-semivariogram between the main and covariates, and then fits and generates a theoretical semivariogram model. Based on this model, the spatial correlation of the covariates is used to perform a weighted estimation of the unsampled areas between beacon nodes. This process expands the sparse point measurement data into a rasterized soil erosion depth distribution matrix with spatial resolution consistent with radar imagery. S504. Erosion Level Classification and Data Output. Based on preset water and soil conservation erosion intensity classification standards, the system numerically segments the rasterized soil erosion depth distribution matrix to generate graded erosion thematic maps. For raster areas where the calculated erosion depth or rate of change exceeds a preset safety threshold, the system automatically extracts their geographic boundary vector coordinates and generates a spatial analysis dataset containing the affected area, average erosion depth, and estimated soil loss, which is then transmitted to the user terminal or visualization screen via a communication interface.

[0026] This invention establishes a monitoring mechanism with air-ground collaboration as the core architecture and orthogonal projection decoupling as the key means by connecting the information flow of physical mechanical displacement, radio frequency coding modulation and spatial statistical inversion. Ultimately, it realizes the closed-loop operation of the entire process of the intelligent sensing and early warning system for watershed soil erosion based on air-ground collaboration.

Claims

1. A watershed soil erosion intelligent sensing and early warning system based on air-space-ground collaborative architecture, characterized in that, include: A synthetic aperture radar observation platform is used to construct an active air-to-ground observation link for downlink radar pulse illumination and uplink echo reception. The variable-depth active coherent beacon includes a deep anchoring unit and a surface floating unit, which is used to capture the vertical displacement difference of the surface. The displacement difference is mapped to the baseband phase modulation amount using a digital gain coefficient set to be greater than 1, and the radar pulse is modulated and forwarded using an orthogonal coding sequence in the slow time domain. The data processing and inversion center is used to construct a matching projection operator to decouple the beacon phase from the radar echo, use a spatial interpolation algorithm to invert and generate a soil erosion depth distribution map of the entire watershed, and compare the distribution map data with a preset safety threshold model. When the threshold is exceeded, a graded early warning signal is generated.

2. The intelligent sensing and early warning system for watershed soil erosion based on air-space-ground coordination as described in claim 1, characterized in that, The variable-depth active coherent beacon also includes a high-precision displacement sensing unit, a digital control unit, and a radio frequency front-end module. The deep anchoring unit is fixed to a non-eroding stable soil layer or bedrock, the surface floating unit is in close contact with the soil surface and settles as the soil is washed away, and the high-precision displacement sensing unit connects the two to measure the vertical displacement difference. The radio frequency front-end module includes a receiving antenna, a transmitting antenna, a low-noise amplifier, a power amplifier, a phase modulator, and a radio frequency circulator or transceiver switch; the radio frequency circulator or transceiver switch is used to isolate the receiving link from the transmitting link; the low-noise amplifier is used to amplify the received radar pulse signal in the pre-stage; the power amplifier is used to perform power gain processing on the phase-modulated signal to achieve active signal forwarding.

3. The intelligent sensing and early warning system for watershed soil erosion based on air-space-ground coordination as described in claim 1, characterized in that, When the variable-depth active coherent beacon performs timing synchronization, it uses a built-in high-frequency clock counter to capture the rising edge of the received radar pulse sequence and lock the pulse repetition frequency. The digital control unit of the variable-depth active coherent beacon has pre-stored inherent circuit processing delay parameters. When predicting the arrival time of the next radar pulse, the inherent circuit processing delay parameters are subtracted from the theoretical arrival time to generate a modulation enable signal with a preset advance, so that the phase modulation action is aligned with the arrival of the radar pulse on the time axis.

4. The intelligent sensing and early warning system for watershed soil erosion based on air-space-ground coordination as described in claim 1, characterized in that, When generating the baseband phase modulation, the variable-depth active coherent beacon acquires the analog signal of the vertical displacement difference through an analog-to-digital converter and uses a digital filtering algorithm to filter out high-frequency vibration components in order to extract the quasi-static displacement value. The variable-depth active coherent beacon calculates the quasi-static displacement value based on the synthetic aperture radar carrier wavelength and the digital gain coefficient. The calculated baseband phase modulation amount is converted into a binary control word through quantization processing, which is used to drive the RF front-end module.

5. The intelligent sensing and early warning system for watershed soil erosion based on air-space-ground coordination as described in claim 1, characterized in that, The modulation of the variable-depth active coherent beacon in the slow time domain specifically includes: Maintain a slow-time index variable corresponding to the pulse number within the synthesis aperture integration time; For each radar pulse, read the symbol value corresponding to the current slow time index in the preset orthogonal coding sequence; Using binary phase shift keying, the encoded phase corresponding to the symbol value is superimposed with the baseband phase modulation amount through digital addition to generate the total modulation phase; The total modulation phase control phase modulator is used to enable the forwarded radio frequency signal to carry orthogonal coding features and erosion physical quantity features on the slow time axis.

6. The intelligent sensing and early warning system for watershed soil erosion based on air-space-ground coordination as described in claim 1, characterized in that, When decoupling signals, the data processing and inversion center first performs range pulse compression on the original echo data, and then constructs a dedicated matching projection operator for each variable-depth active coherent beacon. The matching projection operator includes an orthogonal coding sequence conjugate term for demodulation and a phase correction term for compensating for the Doppler phase history; the phase correction term is constructed based on the instantaneous geometric slant range calculated from the radar platform's precise orbital state vector and the beacon's geographic coordinates.

7. The intelligent sensing and early warning system for watershed soil erosion based on air-space-ground coordination as described in claim 6, characterized in that, When performing coherent integration, the data processing and inversion center uses an interpolation resampling algorithm to extract complex signal values ​​that precisely correspond to the instantaneous geometric slant distance from the discrete data after range-to-pulse compression. The interpolation resampling algorithm is either the Sinc interpolation algorithm or the cubic spline interpolation algorithm; The data processing and inversion center performs a dot product of the extracted complex signal value and the matching projection operator, and sums up all pulse results within the synthetic aperture integration time to achieve coherent gain of the beacon signal and incoherent suppression of background clutter.

8. The intelligent sensing and early warning system for watershed soil erosion based on air-space-ground coordination as described in claim 1, characterized in that, The variable-depth active coherent beacon includes reference beacon nodes deployed in a stable region; The data processing and inversion center extracts the observed phase of the reference beacon node, and uses the least squares method to fit the phase data of the reference beacon node to a planar model or a quadratic surface model to construct an atmospheric phase screen model. The atmospheric phase screen model is used to calculate the estimated atmospheric phase value at each location within the monitoring area, and the estimated atmospheric phase value is subtracted from the observed phase of other variable-depth active coherent beacons to correct for atmospheric delay error.

9. The intelligent sensing and early warning system for watershed soil erosion based on air-space-ground coordination as described in claim 8, characterized in that, When inverting the erosion depth, the data processing and inversion center performs phase unwrapping processing on the corrected net erosion phase in the time dimension. When the net erosion phase difference between the current moment and the previous moment exceeds a preset threshold, the continuous cumulative phase change is restored by adding or subtracting an integer multiple of the period, and the unwrapped phase is converted into physical erosion depth through the inverse operation of the phase mapping model.

10. The intelligent sensing and early warning system for watershed soil erosion based on air-space-ground coordination as described in claim 1, characterized in that, The spatial interpolation algorithm specifically adopts the co-kriging interpolation algorithm; The data processing and inversion center uses the discrete erosion depth calculated by active coherent beacons of varying depths as the main variable, and the surface texture features extracted from synthetic aperture radar amplitude images and the slope data calculated by digital elevation model as the co-variable variables. By calculating the experimental semivariogram of the main variable and the cross semivariogram of the main and covariates, a semivariogram model is fitted and generated. Based on this, a weighted estimate of the unsampled area between beacon nodes is performed to generate rasterized spatial distribution data of erosion depth.

Images