CN119439191A - Differential InSAR system, method, application and readable storage medium - Google Patents

Abstract

The invention provides a differential InSAR system, a differential InSAR method, an application and a readable storage medium. The self-adaptive optical compensation module corrects laser wave front distortion caused by atmospheric turbulence in real time according to a deep learning and wave front detection mechanism, and ensures accurate and stable remote sensing. The innovation point is that the self-adaptive optical performance is improved by combining deep learning, and the phase noise is restrained. The method has profound significance in the fields of topographic mapping, geological disaster monitoring and the like, and can accurately measure the topographic elevation and the tiny displacement of the earth surface, for example, early millimeter-level fluctuation is captured sharply in landslide early warning, so that the method is used for disaster prevention and control.

Description

Differential InSAR system, method, application and readable storage medium

Technical Field

The invention relates to the technical field of radar detection, in particular to a differential InSAR system, a differential InSAR method, a differential InSAR application and a differential InSAR readable storage medium.

Background

In the field of modern radar detection, synthetic Aperture Radar (SAR) technology is evolving continuously, playing a key role in acquiring surface information. The differential InSAR (D-InSAR) technology is particularly prominent, and can accurately monitor the tiny deformation of the earth surface by carrying out interference processing on SAR images acquired at different times, and is widely applied to various fields such as topographic mapping, geological disaster early warning, urban infrastructure monitoring and the like.

However, most current differential InSAR systems face significant challenges in remote sensing. In the prior art represented by CN1329743C, CN101551455B, when the radar is in a long-distance working state, atmospheric turbulence becomes a key factor for restricting performance improvement. The turbulence causes the laser radar wave front to generate complex distortion, and the existing system cannot accurately compensate the wave front distortion due to the lack of an effective correction mechanism, so that obvious phase noise is caused. In geological disaster monitoring scenes such as early stage tiny displacement monitoring of landslide bodies, the noise interference cannot timely and accurately capture changes, the reliability and measurement precision of a system are greatly weakened, the increasing requirements for high-precision, stable and reliable remote surface monitoring are difficult to meet, and the technology is limited to be deeply applied and expanded in key fields such as deformation monitoring of large-scale water conservancy facilities, prevention and control of geological disasters in remote mountain areas and the like.

Disclosure of Invention

The embodiment of the invention provides a differential InSAR system, a differential InSAR method, an application and a readable storage medium, aiming at the problems of phase noise and the like caused by turbulence in the use of a long-distance D-InSAR in the prior art.

The core technology of the invention mainly improves the correction bandwidth of the self-adaptive optical system by a deep learning technology, and solves the phase noise problem caused by turbulent interference of laser radar wave-front information by the self-adaptive optical system.

In a first aspect, the present invention provides a differential InSAR system comprising:

a laser emitting module for generating and distributing a laser beam;

the radar detection module is used for detecting a target and acquiring and processing signals;

The self-adaptive optical compensation module is used for correcting phase distortion caused by atmospheric turbulence;

The radar detection module comprises two paths of detection branches which are symmetrical in structure and work cooperatively, the two paths of detection branches respectively detect targets at the same time and acquire and process echo signals so as to realize a differential detection function and improve the reliability and the accuracy of the system, and the self-adaptive optical compensation module acquires and compensates laser wavefront distortion caused by atmospheric turbulence in real time according to a deep learning technology and a wavefront detection processing mechanism so as to ensure the stable and accurate detection of the differential InSAR system in long-distance remote sensing.

Further, the adaptive optics compensation module comprises an adaptive optics system laser, a digital micromirror device and a wavefront sensor;

the self-adaptive optical system laser emits measurement laser when the tunable laser of the laser emitting module works, scattered wave fronts are sequentially received by the digital micro-mirror device, compressed and encoded and the wave front sensor receives the laser in one exposure time after the laser is influenced by atmospheric turbulence;

reconstructing continuous images from the one-time exposure result of the wavefront sensor by means of the trained deep learning network model so as to improve the frame rate and the closed-loop bandwidth;

and calculating ideal wavefront deviation according to wavefront reconstruction, determining conjugate wavefront execution quantity, and driving a first deformable mirror and a second deformable mirror of a radar detection module to compensate the radar to receive the laser wavefront so as to achieve the purpose of turbulence resistance.

When the radar detection module works in a single navigation mode, the calibration light of the adaptive optical compensation module is positioned at the radar transmitting end, atmospheric turbulence phase distortion is detected in real time, the wavefront sensor and the deformable mirror are arranged in the radar receiving end, the receiving antenna receives target reflection echoes, the wavefront sensor firstly measures the distortion calibration light, the compensation coefficient is obtained through an algorithm, the deformable mirror is sent to remove distortion, and the correction echoes are received by the radar receiving end.

Further, the laser emitting module comprises a tunable laser which divides laser light into two paths by a beam splitter into a radar detection module.

Further, the radar detection module comprises a master radar and a slave radar;

two paths of laser emitted by the laser emitting module are collimated by a collimator, polarized by a polarizer, aperture filtered and polarized by a half wave plate in respective branches of the main radar and the auxiliary radar, the polarized beam splitter divides the local oscillation laser into local oscillation laser and emergent laser, the local oscillation laser enters a receiving light path through a reflecting mirror and a quarter wave plate and the polarized state is converted, the emergent laser irradiates a target, reflected light corrects the wave front phase through a deformable mirror and is mixed with the local oscillation laser, an interference pattern and the target alignment condition are obtained by observing by a charge coupled device, and the rest is detected by an avalanche photodiode after the aperture filtered for time sequence processing.

Further, the secondary radar comprises a first collimator, a first polarizer, a first aperture, a first half-wave plate, a first polarizing beam splitter, a first quarter-wave plate, a second quarter-wave plate, a first reflecting mirror, a second polarizing beam splitter, a first charge coupled device, a second aperture, a first avalanche photodiode, a first deformable mirror;

The main radar comprises a second collimator, a second polarizer, a third small hole, a second half wave plate, a third polarizing beam splitter, a third quarter wave plate, a fourth quarter wave plate, a second reflecting mirror, a fourth polarizing beam splitter, a second charge coupled device, a fourth small hole, a second avalanche photodiode and a second deformable mirror.

Further, the adaptive optical compensation module adopts EFFICIENTSCI deep learning network model according to deep learning technology.

In a second aspect, the present invention provides a control method of a differential InSAR system, including:

s00, generating and distributing a laser beam to enter a radar detection module by a laser emission module;

s10, two paths of detection branches of the radar detection module respectively detect targets at the same time and acquire and process echo signals so as to realize a differential detection function;

S20, the self-adaptive optical compensation module acquires and compensates laser wavefront distortion caused by atmospheric turbulence in real time according to a deep learning technology and a wavefront detection processing mechanism.

In a fourth aspect, the present invention provides a readable storage medium having stored therein a computer program comprising program code for controlling a process to execute a process, the process comprising a control method according to the above.

The main contributions and innovation points of the invention are as follows:

1. In a long-distance D-InSAR use scene, the prior art cannot correct wavefront distortion, so that the phase noise of a radar system is obvious, and the measurement accuracy is seriously interfered. The invention can accurately correct the phase distortion caused by turbulence in the observation area in real time by means of the subtle cooperation of the deep learning technology and the self-adaptive optical system. The influence of atmospheric turbulence on the laser wavefront is monitored in real time, and the deformation mirror is driven by an advanced algorithm to implement displacement compensation, so that phase noise is effectively restrained. When the topography is mapped, the technology greatly improves the measurement precision of the topography elevation, can sharply catch the tiny displacement change of the earth surface in the field of geological disaster monitoring, such as early millimeter displacement of landslide, strives for key time for disaster early warning, and greatly enhances the detection reliability and accuracy of the system in complex atmospheric environment.

2. The invention innovatively introduces a deep learning technology to optimize the self-adaptive optical system. And processing the wavefront sensor data by using EFFICIENTSCI model, realizing reconstruction of continuous images from single exposure, greatly increasing the frame rate of the wavefront sensor, and further expanding the closed-loop bandwidth of the self-adaptive system. Compared with the traditional self-adaptive optical system, the invention can respond to the change of the atmospheric turbulence more sensitively and compensate the disturbance of the turbulence which changes rapidly in time. In urban infrastructure monitoring, in the face of complex airflow interference among high-rise building groups, the system is used for stably and accurately monitoring bridges and building micro deformation, and provides firm technical support for urban safety operation and maintenance, and the application range and application scene of the system are powerfully widened.

3. The system has the advantages of enhancing the overall stability and adaptability of the system, fusing the unique architecture and technology, and endowing the system with strong stability and adaptability. Under various complex climates and geographic conditions, high-precision detection performance is always maintained. Whether the high-altitude area is aggravated by strong wind and low-temperature atmospheric turbulence or the coastal area is challenged by humidity and salt erosion, the system can be self-adaptively adjusted, and the system is stable and reliable to operate. In the long-term monitoring task, such as deformation monitoring of large-scale water conservancy facilities, analysis data are accurately collected continuously for many years, a coherent accurate basis is provided for facility safety evaluation and maintenance planning, interference and influence of environmental factors on system performance are effectively reduced, the service life of the system is prolonged, the full life cycle cost performance is improved, and key performances are played in the fields of global geographic information collection, infrastructure safety assurance and the like.

The details of one or more embodiments of the invention are set forth in the accompanying drawings and the description below to provide a more thorough understanding of the other features, objects, and advantages of the invention.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention and do not constitute a limitation on the invention. In the drawings:

FIG. 1 is a diagram of the optical path architecture of a radar transceiver of a differential InSAR system in accordance with an embodiment of the present invention;

FIG. 2 is a graph of the coordinate relationship of a radar airborne platform to a target in accordance with an embodiment of the invention;

FIG. 3 is a diagram of a radar transmit and receive aperture distribution pattern according to an embodiment of the present invention;

FIG. 4 is a diagram of the spatial relationship of the interference portion of a radar transceiver to a target in accordance with an embodiment of the present invention;

FIG. 5 is a diagram of the spatial coordinate relationship of a radar to a target according to an embodiment of the present invention;

FIG. 6 is a diagram of a radar transmit-receive caliber layout versus target coordinates in accordance with an embodiment of the present invention;

fig. 7 is a system block diagram of a differential InSAR system in accordance with an embodiment of the present invention.

In the figure, 1, a tunable laser, 2, a Slave radar part, 3, a Master radar part, 4, a target, 5, a beam splitter, 6-1, an adaptive optical system laser, 6-2, a wavefront sensor and 6-3, a digital micromirror device;

2-1, a first collimator, 2-2, a first polarizer, 2-3, a first aperture, 2-4, a first half wave plate, 2-5, a first polarization beam splitter, 2-6, a first quarter wave plate, 2-7, a second quarter wave plate, 2-8, a first reflecting mirror, 2-9, a second polarization beam splitter, 2-10, a first charge coupled device, 2-11, a second aperture, 2-12, a first avalanche photodiode, 2-13, a first deformable mirror;

3-1, a second collimator, 3-2, a second polarizer, 3-3, a third aperture, 3-4, a second half wave plate, 3-5, a third polarization beam splitter, 3-6, a third quarter wave plate, 3-7, a fourth quarter wave plate, 3-8, a second reflecting mirror, 3-9, a fourth polarization beam splitter, 3-10, a second charge coupled device, 3-11, a fourth aperture, 3-12, a second avalanche photodiode, 3-13, a second deformable mirror.

Detailed Description

Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numbers in different drawings refer to the same or similar elements, unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with one or more embodiments of the present specification. Rather, they are merely examples of apparatus and methods consistent with aspects of one or more embodiments of the present description as detailed in the accompanying claims.

It should be noted that in other embodiments, the steps of the corresponding method are not necessarily performed in the order shown and described in this specification. In some other embodiments, the method may include more or fewer steps than described in this specification. Furthermore, a single step described in this specification may be described as being split into multiple steps in other embodiments, while multiple steps described in this specification may be described as being combined into a single step in other embodiments.

Example 1

The present invention aims to propose a differential InSAR system, in particular with reference to fig. 1-7, comprising:

The tunable laser comprises a tunable laser 1, a 1-minute 2 optical fiber beam splitter 5, a Slave radar part 2, a Master radar part 3, a target 4, an adaptive optics system laser 6-1, a wavefront sensor 6-2 and a digital micro-mirror device 6-3;

Wherein the Slave radar section 2 (Slave radar) further includes a first collimator 2-1, a first polarizer 2-2, a first aperture 2-3, a first half-wave plate 2-4, a first polarizing beam splitter 2-5, a first quarter-wave plate 2-6, a second quarter-wave plate 2-7, a first reflecting mirror 2-8, a second polarizing beam splitter 2-9, a first charge coupled device 2-10, a second aperture 2-11, a first avalanche photodiode 2-12, a first deformable mirror 2-13;

The Master radar section 3 (main radar) further includes a second collimator 3-1, a second polarizer 3-2, a third aperture 3-3, a second half wave plate 3-4, a third polarizing beam splitter 3-5, a third quarter wave plate 3-6, a fourth quarter wave plate 3-7, a second reflecting mirror 3-8, a fourth polarizing beam splitter 3-9, a second charge coupled device 3-10, a fourth aperture 3-11, a second avalanche photodiode 3-12, and a second deformable mirror 3-13.

In this embodiment, the tunable laser 1 divides the laser into two paths through the beam splitter 5, one path enters the Slave radar section 2, and the other path enters the Master radar section 3, and the two paths are identical, so that the targets are detected simultaneously.

After being collimated by a first collimator 2-1 and a second collimator 3-1, laser emitted by an optical fiber is polarized by a first polarizer 2-2 and a second polarizer 3-2, the polarized laser beam is rotated to a proper angle by a first half wave plate 2-4 and a second half wave plate 3-4 through a first small hole 2-3 and a third small hole 3-3, the power ratio of local oscillation laser and echo signals of a Master radar part and a Slave radar part is controlled, the first polarization beam splitter 2-5 and the third polarization beam splitter 3-5 divide the light beam into local oscillation laser and emitted laser, the local oscillation laser is reflected by a first quarter wave plate 2-7 and a second quarter wave plate 3-7 and a first reflecting mirror 2-8 and a second reflecting mirror 3-8 to enter a receiving light path, the polarization state is changed from vertical polarization to horizontal polarization, the emergent laser irradiates the surface of a target 4 after passing through a first quarter wave plate 2-6 and a third quarter wave plate 3-6, the backward scattered laser corrects the distorted wave front phase through a first deformable mirror 2-13 and a second deformable mirror 3-13, is mixed with local oscillation laser after passing through a second polarization beam splitter 2-9 and a fourth polarization beam splitter 3-9, a part of the mixed wave front phase is observed by a first charge coupled device 2-10 and a second charge coupled device 3-10 and is used for obtaining an interference pattern, the alignment condition of the target is checked, a part of the mixed wave front phase is detected by a first avalanche photodiode 2-12 and a first avalanche photodiode 3-12 after passing through a second small hole 2-11 and a fourth small hole 3-11 for low-pass filtering, for synthetic aperture treatment.

The adaptive optics system laser 6-1, the wavefront sensor 6-2 and the digital micromirror device 6-3 form an adaptive optics module (adaptive optics system), as shown in fig. 7, the calibration light of the adaptive optics system (AO) is located in the radar transmitting part and used for detecting the phase distortion caused by the atmospheric turbulence in real time, the receiving antenna receives the echo information reflected from the target, the wavefront detector detects the distorted calibration light, the wavefront reconstruction algorithm calculates the compensation coefficient, and sends the compensation coefficient to the deformable mirror to remove the phase distortion caused by the turbulence, and finally the corrected echo information is received by the radar receiver. The measuring process of the adaptive optics system is as follows:

While the tunable laser 1 works, the self-adaptive optical system laser 6-1 also emits measurement laser, after the influence of atmospheric turbulence, scattered wave front is firstly received by the digital micro mirror device 6-3, N times of compression coding are carried out, then the scattered wave front is received by the wave front sensor 6-2 within one exposure time, N times of continuous images are rebuilt from one exposure result of the wave front sensor 6-2 through a trained EFFICIENTSCI depth learning network model, so that the frame rate of the wave front sensor 6-2 is improved, the frame rate of the self-adaptive system is improved, the integral closed loop bandwidth of the self-adaptive system is improved, the influence of the atmosphere on the wave front of the laser at the current time is obtained, the deviation condition of the ideal wave front is calculated through wave front reconstruction, the execution quantity required by the conjugate wave front is calculated, and then the execution quantity is transmitted to the first deformable mirror 2-13 and the second deformable mirror 3-13 to execute corresponding displacement compensation, and accordingly the laser wave front of a radar receiving part is compensated, and the purpose of turbulence resistance is achieved.

Wherein the avalanche photodiode is APD (Avalanche Photo Diode), which is a photodetector with internal gain. When photons are incident into the absorption region of the photoelectric conversion device to generate electron-hole pairs, the electrons are accelerated to collide with lattice atoms by the high electric field region to trigger an avalanche multiplication effect to generate more carriers, so that the photoelectric current is amplified, the detection sensitivity to weak optical signals is improved, and the weak optical signals can be converted into obvious electric signals. In the invention, the optical signal after being subjected to small-hole filtration is detected by the first avalanche photodiode 2-12 and the second avalanche photodiode 3-12 in time series, and the time series data has great significance for subsequent synthetic aperture processing.

The digital micromirror device is DMD (Digital Micromirror Device), which is a microelectromechanical system (MEMS) based light modulation element. The core of the system is a plurality of tiny mirror arrays capable of independently controlling the inclination angle, and the direction, the intensity and the phase of reflected light are accurately modulated by controlling the mirrors to switch in the on state and the off state or deflect at different angles, so that the digital control of the spatial distribution and the intensity of the optical signals is realized. In the present invention, the DMD plays a key role. The atmospheric turbulence causes the laser wavefront to distort, the DMD receives the scattered wavefront and performs N compression encodings, enabling the wavefront sensor 6-2 to capture multiple wavefront information at a single exposure.

The Charge-Coupled Device is a CCD (Charge-Coupled Device), which is a semiconductor Device for converting an optical image into a digital signal. The core is that the pixel matrix is composed of a series of closely arranged photosensitive elements. When light irradiates the photosensitive element, photons excite to generate electron-hole pairs, and electrons are collected and stored in potential wells corresponding to pixels to form charge packets proportional to the illumination intensity. Under the drive of clock pulse, these charge packets are transferred in turn according to a specific sequence, so as to implement the output of one-dimensional serial electric signal from two-dimensional pixel array, and the electric signal is converted into digital image data by the following processes of amplification, sampling, quantization and coding. In the present invention, two CCDs assume a critical task. In the radar detection process, reflected light after the outgoing laser irradiates the target is processed and mixed with local oscillation laser, and part of mixed light is received by a CCD.

Preferably, in fig. 3B is the center distance between the transmitting antenna and the receiving antenna, d is the center distance between the receiving antenna and the transmitting antenna, and s is the center distance between the transmitting antenna and the receiving antenna.

In this embodiment, the radar transmission process is as shown in FIGS. 4-6, wherein FIG. 4Main transmitting antenna and target connection line the direction is at an included angle with the vertical direction,From the angle between the direction of the connecting line of the transmitting antenna and the target and the vertical direction,Is the connection distance between the main transmitting antenna and the target,Is the distance between the transmitting antenna and the target, the height of the H radar from the ground, the height of the H target,And the target height is at a coordinate corresponding to the z axis. FIG. 5 shows a coordinate rotation process, in which a new z-axis is centered on a transmitting antenna, a new direction of a line connecting the transmitting antenna center and a target is centered on the new z-axis, a direction pointing to the ground perpendicular to the new z-axis is a new x-axis, the new coordinate system x 'oz' is translated and rotated relative to the original xoz coordinate system, the translation vector is [ B/2,0, H ], and the rotation angle is. FIG. 6 shows a schematic diagram of a computer systemFor the sampling interval of the radar system,Representing the sequence of samples of the nth time,Is the diagonal distance of the main receiving antenna to the target.

Assuming that two beams of coherent light are emitted, interference occurs at a target point, and a target p isThe moment is in the plane of the transmitting antenna. Let MT (master transmitter), ST (slave transmitter) each emit gaussian beam and omit constant term and unimportant amplitude term, it is possible to obtain:

x, y and z are the spatial coordinates of the beam, the first exp term is the phase delay caused by the propagation of the beam, and the second exp term is the secondary phase caused by the Gaussian beam.

Let aircraft speed beThe step time interval isThe step number of the azimuth direction is n, thenAt the moment, the center coordinate of the transmitting antenna is. The diffraction distance of the gaussian beam is in fact the distance of the transmitting antenna to the target, which is related to the slip angle θ. When the target is sufficiently far away from the object,。

The center coordinates of the transmitting antenna areMT center coordinates areST center coordinates are。

Assuming that the elevation information of the target is h, the coordinates of the upper end of the target point can be expressed as:

the light field that can get the Gaussian light emitted by ST to reach the target point is:

The light field of the MT-emitted gaussian light reaching the target point is:

The interference field formed at the target point is:

The light intensity formed by the target point is used as a point target (point light source), the light intensity of interference fringes is received by the MR after back scattering to the receiving antenna. Order the The light field of the scattering point target to the receiving surface can be expressed as:

The coordinates of the plane where the scattering points are located, and x and y are the coordinates of the receiving surface.

Since the tilt from the target point to the MR can be accurately measured during the receiving process,Is the diffraction distance of the transmitting radar to the target, so the amplitude (the mode of complex amplitude) received by the receiving antenna MR is actually a distanceAnd (3) the decay value of the (C) is obtained by classifying the irrelevant term into a constant term C:

For the reflectivity of the object to be achieved, For the radar wavelength to be of the same wavelength,To obtain the optical path difference for the target to MR by inverse solutionThereby can pass accuratelyObtainingI.e. by means of the information of the transmitter we can obtain elevation information of the target.

The radar platform receiving apertures are deployed along the y direction, the interval of the receiving apertures is d, the two apertures continuously scan the same area, and tiny target area change information is extracted by comparing measurement data at different times. The processing flow comprises the following steps:

S1, data preprocessing:

Denoising and filtering, namely denoising the laser echo signal (such as wavelet transformation, median filtering and Gaussian filtering), filtering out environmental noise and interference signals and improving the signal quality.

Wherein, wavelet transformation denoising step:

1. the wavelet basis and the number of decomposition layers are selected by selecting a mother wavelet function (e.g., daubechies wavelet, symlets wavelet, etc.) that is suitable for the signal characteristics. The number of layers of the decomposition is determined, typically based on the frequency characteristics of the signal and the detail requirements.

2. Wavelet decomposition-the original laser echo signal or image is subjected to a multi-layer wavelet decomposition into sub-bands of different frequencies (low and high frequencies).

3. Thresholding, namely, applying a threshold to the wavelet coefficients of the high-frequency sub-bands to remove noise. Common thresholding methods include:

and a hard threshold (HardThresholding) in which coefficients whose absolute value is lower than the threshold are set to zero, while coefficients higher than the threshold are kept unchanged.

And a soft threshold (SoftThresholding) in which the absolute value of the coefficient below the threshold is zero and the coefficient above the threshold is subtracted from the threshold value.

The threshold value may be selected based on statistical characteristics (such as noise standard deviation) or empirical methods.

4. Wavelet reconstruction, which is to reconstruct a denoised signal or image by inverse wavelet transform using the thresholded coefficients.

5. Post-processing (optional) further smoothing or enhancement processing is performed as needed to optimize the denoising effect.

Wherein, the median filtering noise reduction step

1. The filter window size is selected by determining the size of the sliding window (e.g., 3 x 3, 5 x 5, etc.), the window size being selected based on the noise type and signal characteristics.

2. Sliding window traversal slides a window of a selected size pixel by pixel (or point by point) on a signal or image.

3. And calculating a median value, namely sequencing all pixel values in the window, and taking the middle value as a new value of the central pixel of the current window.

4. And replacing the pixel value, namely assigning the calculated median value to the pixel in the center of the window to finish one-time filtering operation.

5. The traversal is repeated, repeating the above process for the entire signal or image until all pixels are processed.

Wherein, gaussian filtering noise reduction step:

1. the size and standard deviation (σ) of the gaussian kernel are chosen:

the window size (e.g., 3 x 3, 5 x 5, etc.) and standard deviation sigma of the gaussian filter are determined, and the sigma value determines the smoothness of the filtering.

2. Constructing a Gaussian kernel:

And calculating the weight of the Gaussian function at each window position according to the selected window size and sigma value, and generating a Gaussian kernel matrix.

The sum of gaussian kernels is ensured to be 1 to keep the image brightness unchanged.

3. Image convolution:

The gaussian kernel is convolved with the original signal or image, i.e., the gaussian kernel is slid over each pixel of the signal, and a weighted average is calculated.

For each pixel location, the product of the surrounding pixels and the gaussian kernel weight is calculated as the new pixel value.

4. Boundary processing:

the pixels of the edge area of the signal or image are processed, and common methods include zero padding, mirror padding or edge replication to ensure that the filter is applied at the edges.

5. Outputting a denoising result:

after the convolution operation is completed, a smooth denoised signal or image is obtained.

6. Time synchronization, which is to ensure the time synchronization of laser echo data and navigation data (IMU, GPS) and prepare for subsequent track correction and data alignment.

The laser echo data and the navigation data are matched and aligned according to a time stamp, for example, using a time stamp recorded during data acquisition, based on the alignment (TIMESTAMPALIGNMENT) of the time stamps. The operation steps are as follows:

1. extracting time stamp, namely extracting respective time stamp information from the laser echo data and the navigation data.

2. And matching the data points in the two data streams according to the time stamps.

3. Interpolation processing-if the time steps of two data streams are not identical, the time step of one of the data streams is adjusted to be identical to the other using interpolation methods (e.g., linear interpolation, spline interpolation).

4. And (3) aligning the data, namely aligning the adjusted data according to a time sequence, and ensuring the corresponding time of each data point to be consistent.

S2, track correction and alignment:

5. track reconstruction, namely reconstructing the accurate motion track of the sensor according to the IMU and GPS data.

The Kalman filtering in Kalman filtering track reconstruction is a recursive algorithm, and is widely applied to sensor data fusion and track estimation. It is applicable to linear systems, but there are also extended versions (e.g. extended kalman filters, EKFs) for nonlinear systems. The operation steps are as follows:

1. state vectors, which typically include parameters such as position, velocity, acceleration, and the like, and observation vectors are defined. For example, x= [ x, y, z, vx, vy, vz ] tx= [ x, y, z, vx, vy, vz ] T. The observation vector includes position data provided by the GPS and acceleration, angular velocity data provided by the IMU.

2. Establishing a state transition model:

how the system transitions from one state to the next is described, typically based on kinematic equations. For example, the location update may be based on velocity, and the velocity update based on acceleration.

3. Establishing an observation model:

describing how observation data is generated from the state vector. The GPS location data is mapped directly to the location portion in the state vector, and the IMU data is used to predict acceleration and angular velocity.

4. Initializing a filter:

setting an initial state estimation and a covariance matrix.

5. Recursive prediction and update:

and predicting the next state and covariance based on the state transition model.

And updating, namely updating the state estimation and the covariance by combining the new observed data.

6. Output track:

After multiple iterations, the filter outputs a smooth and accurate trajectory estimate.

7. And (3) data alignment, namely spatially aligning scanning data at different time or different positions, and eliminating errors caused by platform movement.

Among these, for example, the iterative closest point algorithm (ITERATIVE CLOSEST POINT, ICP) is an algorithm widely used for point cloud data alignment, finding the best rigid transformations (rotation and translation) between two point clouds by iterative optimization so that they coincide as much as possible. The operation steps are as follows:

1. Initial alignment-initial alignment of two point clouds, typically based on rough sensor pose estimation.

2. And (3) nearest point matching, namely, for each point in the source point cloud, finding the nearest point in the target point cloud as a corresponding point.

3. Transformation estimation-computing the best rigid transformation (rotation and translation) to transform the source point cloud to the target point cloud is typically accomplished by a least squares method.

4. And (3) applying the transformation, namely applying the calculated transformation to the source point cloud.

5. And (3) convergence judgment, namely checking whether the transformation is converged (if the transformation amount is smaller than a threshold value), and if the transformation is not converged, repeating the steps (2-4).

S3, synthetic aperture treatment:

6. and (3) coherent accumulation, namely superposing the data of multiple scans, and improving the spatial resolution by utilizing a synthetic aperture technology.

Wherein, for example, coherent accumulation and matched filtering (Coherent Accumulationwith MATCHED FILTERING) are a signal processing technology, and detection and enhancement of signals are realized through matching with a known signal template. In coherent accumulation, the coherence of the target signal can be further enhanced by combining matched filtering, and noise is suppressed. The operation steps are as follows:

1. Template selection, selecting or generating a template signal of a matched filter based on characteristics of the intended target (e.g., frequency, phase, etc.).

2. Filtering application-a matched filter is applied to each data segment to be accumulated, extracting the portion that matches the template signal.

3. Coherent accumulation, namely, carrying out coherent accumulation on the data subjected to matched filtering, and generally adopting a phase alignment and weighted superposition method.

4. And the signal enhancement is to obviously enhance the coherence of the target signal and inhibit random noise and interference through matched filtering and coherent accumulation.

5. And phase correction, namely, for coherent measurement, performing phase correction to eliminate phase offset caused by platform motion and environmental factors.

S4, differential calculation:

And differential matching, namely comparing the high-resolution point clouds or image data at different time or different positions point by point to calculate differential information.

And detecting small changes of the target area, such as displacement, deformation, terrain change and the like, based on the differential data.

Among them, for example, time series analysis (Time-SERIES ANALYSIS) is to analyze multi-phase differential data to identify a trend and an abnormality of Time variation, thereby realizing variation detection. The operation steps are as follows:

1. and collecting multi-time phase data, namely acquiring point clouds or image data of a plurality of time points and generating a multi-time phase differential interference image.

2. Time series modeling, namely establishing a time series model such as an autoregressive model (AR), a moving average Model (MA), an autoregressive moving average model (ARMA) and the like for each monitoring point.

3. And (3) trend and anomaly detection, namely analyzing trend, periodicity and anomaly points in the time sequence, and identifying the moment and the area of anomaly change.

4. And marking a change area, namely marking a significantly changed area according to the deviation of model prediction and actual observation.

5. And (3) post-processing, namely carrying out noise suppression and connectivity optimization of a change area by combining spatial information.

S5, error correction and optimization:

geometric error correction, namely correcting geometric errors caused by systematic errors, vibration of a moving platform and the like.

And 3, precision optimization, namely, optimizing through an algorithm (least square method), improving the precision and reliability of differential measurement, and reducing error propagation. The least squares method (Least Squares Method) is a classical parameter estimation method that optimizes parameter estimation by minimizing the sum of squares of errors between observations and model predictions. This is particularly important in trajectory correction, baseline estimation, and phase correction. Wherein, the operation steps are as follows:

1. And (3) establishing an observation equation, namely connecting actual observation data with a theoretical model to form a linear or nonlinear observation equation.

2. An error function is defined, which is typically the sum of squares of the observed error and the model prediction error.

3. Solving the optimization problem, namely iteratively adjusting model parameters by using optimization technologies such as a Gaussian-Newton method, a Levenberg-Marquardt algorithm and the like, so as to minimize an error function.

4. And (3) evaluating the result, namely analyzing the optimized parameter estimated value and the confidence interval thereof, and ensuring the rationality and the accuracy of the result.

S6, data fusion and imaging:

And generating a three-dimensional point cloud, namely fusing the differential data with the synthetic aperture imaging data to generate a high-precision three-dimensional point cloud model.

Such as point cloud generation (DEM-BasedPoint Cloud Generation) based on an elevation model. By combining Digital Elevation Model (DEM) and DINSAR data, the terrain information provided by the DEM is fused with the interference phase data to generate a high-precision three-dimensional point cloud. The method is suitable for the data scene with the existing elevation reference. The operation steps are as follows:

Dem acquisition and preprocessing, i.e., acquiring a high resolution digital elevation model of a target area and preprocessing (e.g., smoothing, filtering).

2. And generating an interference pattern and highly disentangling, namely generating the interference pattern, and highly disentangling by taking the DEM as a reference to recover the actual height of the ground object.

3. And (3) constructing a point cloud, namely combining the unwrapped height information with geographic coordinates to generate a three-dimensional point cloud.

4. And (3) data fusion, namely fusing the generated three-dimensional point cloud with other data sources (such as optical images and LiDAR data) and improving the details and the precision of the point cloud.

And (3) changing the monitoring image, namely generating a monitoring image for displaying the change area, and intuitively displaying the tiny change of the target area. For example, threshold segmentation is the most basic method of detecting a change, and a change region is distinguished from a non-change region by setting one or more thresholds. From the segmentation result, a binarized change monitor image is generated, wherein the change region is typically highlighted in a specific color or mark. The operation steps are as follows:

1. And (3) generating differential data, namely generating a differential interference map or a displacement map through a differential matching step (S4).

2. Setting threshold value, namely setting one or more threshold values according to the statistical characteristics (such as mean value and standard deviation) of the data, and distinguishing the changed area from the unchanged area.

3. Applying a threshold value by comparing the differential data to the threshold value, the pels that exceed the threshold value being marked as a change region.

4. A binarized image is generated by converting the marking result into a binary image, the changed areas are usually represented by white or red, and the unchanged areas are represented by black or other colors.

5. And (3) post-processing, namely performing morphological operations (such as expansion and corrosion) to remove isolated noise points, filling small holes and improving the consistency and accuracy of a change area.

Thus, the D-InSAR differential synthetic aperture laser radar realizes high-resolution and high-precision three-dimensional imaging and change monitoring by combining a synthetic aperture technology, an interference technology and a differential measurement method. The implementation process involves the application of precise integration of system design, accurate acquisition of multi-phase or multi-view data, efficient data processing and differential algorithms.

Example two

Based on the same conception, the invention also provides a control method of the differential InSAR system, which comprises the following steps:

s00, generating and distributing a laser beam to enter a radar detection module by a laser emission module;

s10, two paths of detection branches of the radar detection module respectively detect targets at the same time and acquire and process echo signals so as to realize a differential detection function;

S20, the self-adaptive optical compensation module acquires and compensates laser wavefront distortion caused by atmospheric turbulence in real time according to a deep learning technology and a wavefront detection processing mechanism.

In this embodiment, the deep learning technique uses EFFICIENTSCI models, and the training process of the models is as follows:

1) Data collection and preprocessing

And collecting data, namely collecting samples from the actual measurement data of the wavefront sensors under a large number of different atmospheric turbulence conditions, covering various intensity turbulence scenes and corresponding wavefront distortion information, constructing rich and various data sets, faithfully recording the wavefront changes under different moments, climates and geographic environments, and providing comprehensive materials for model learning turbulence characteristic rules.

Preprocessing, namely denoising and normalizing the original data. The filtering algorithm is adopted to remove artifacts and fluctuation caused by measurement noise and environmental interference, the data characteristic value is contracted to a specific range, the data quality and stability are improved, the model focusing learning of wavefront distortion essential characteristics is facilitated, the misleading of learning direction due to data magnitude and distribution difference is avoided, and the model training is ensured to be efficient and accurate.

2) Model architecture construction

The multi-layer convolutional neural network architecture is designed as a main body, and residual connection and attention mechanism optimization are integrated. The method comprises the steps of automatically extracting local features of an image by a convolution layer, expanding receptive fields by multi-layer stacking, capturing a complex space structure of wavefront distortion, relieving gradient disappearance by residual connection, assisting efficient training of a deep network, focusing key regions and features of the image by an attention mechanism, enhancing important feature learning weights, such as highlighting wavefront details of a turbulent strong interference region, and improving model feature extraction precision and characterization capability.

3) Training algorithm selection and parameter setting

And (3) training an algorithm, namely optimizing model parameters by adopting a random gradient descent variant (such as Adam), flexibly adjusting super parameters such as learning rate, momentum and the like according to the data scale, computing resources and convergence speed requirements, balancing training efficiency and model precision, enabling the model to quickly approach an optimal solution, avoiding sinking into local extremum, and efficiently learning the mapping relation between wavefront distortion and correction strategy.

And (3) setting parameters, namely reasonably dividing the training set, the verification set and the ratio of the test set (common 7:2:1), and monitoring the training by using the verification set to prevent overfitting. And determining the batch size and the number of training wheels according to the characteristics of the data set and the hardware performance, guaranteeing the generalization capability and stability of the model, and improving the wave front distortion performance in unknown turbulence scenes through multiple rounds of iterative optimization.

4) Model evaluation and optimization

And evaluating indexes, namely selecting indexes such as Mean Square Error (MSE), structural Similarity Index (SSIM) and the like to quantify the similarity between the model reconstructed image and the real image. The MSE measures the square mean value of pixel value errors to reflect the overall deviation, SSIM evaluates the perceived quality from the multi-dimensions of brightness, contrast and structure, and the comprehensive index accurately evaluates the model performance and indicates the direction for optimization.

And (3) optimizing strategy, namely fine-tuning network architecture, parameters or data enhancement technology according to the evaluation result. If the amplified data enriches sample diversity, improves model complexity balance fitting capability, improves the accuracy and reliability of wavefront distortion under complex turbulence condition by the model processing through multi-round iteration optimization, and meets the requirement of high-accuracy self-adaptive optical compensation of the radar system.

Example III

The present embodiment provides for applying the embodiment to an aircraft, such as a civil aircraft. As shown in fig. 2, the radar is located in the flight direction coordinate system of the aircraft. The radar system works in a single-navigation mode (a platform where the radar is flying above a target once), the radar layout is shown in fig. 3, the calibration light of the adaptive optics system is located at the radar transmitting end (a main transmitter and a slave transmitter), and the wavefront sensor 6-2 and the deformable mirror are located in the radar receiving end (the main receiver and the slave transmitter).

The radar transmitting end comprises a tunable laser 1, a one-to-two optical fiber beam splitter 5, a first collimator 2-1, a second collimator 3-1, a first polarizer 2-2, a second polarizer 3-2, a first half wave plate 2-4, a second half wave plate 3-4, a first polarization beam splitter 2-5, a third polarization beam splitter 3-5, a first quarter wave plate 2-6, a third quarter wave plate 3-6, a first reflecting mirror 2-8 and a second reflecting mirror 3-8.

The radar receiving end comprises a first deformable mirror 2-13, a second deformable mirror 3-13, a second polarization beam splitter 2-9, a fourth polarization beam splitter 3-9, a first charge coupling device 2-10, a second charge coupling device 3-10, a second small hole 2-11, a fourth small hole 3-11, a first avalanche photodiode 2-12 and a second avalanche photodiode 3-12.

Adopting a point target model, wherein the coordinates of a target point are as followsWhen (when)The time instant target p is located in the transmit antenna normal plane (x-z plane),At that point, the aircraft steps forward a distance. The transmitting antennas are located in a normal plane (x-z plane) and the receiving antennas are arranged in azimuth.

Example IV

The present embodiment also provides a readable storage medium having stored therein a computer program including program code for controlling a process to execute the process, the process including the control method according to the second embodiment.

It should be noted that, specific examples in this embodiment may refer to examples described in the foregoing embodiments and alternative implementations, and this embodiment is not repeated herein.

In general, the various embodiments may be implemented in hardware or special purpose circuits, software, logic or any combination thereof. Some aspects of the invention may be implemented in hardware, while other aspects may be implemented in firmware or software which may be executed by a controller, microprocessor or other computing device, although the invention is not limited thereto. While various aspects of the invention may be illustrated and described as block diagrams, flow charts, or using some other pictorial representation, it is well understood that these blocks, apparatus, systems, techniques or methods described herein may be implemented in, as non-limiting examples, hardware, software, firmware, special purpose circuits or logic, general purpose hardware or controller or other computing devices, or some combination thereof.

Embodiments of the invention may be implemented by computer software executable by a data processor of a mobile device, such as in a processor entity, or by hardware, or by a combination of software and hardware. Computer software or programs (also referred to as program products) including software routines, applets, and/or macros can be stored in any apparatus-readable data storage medium and they include program instructions for performing particular tasks. The computer program product may include one or more computer-executable components configured to perform embodiments when the program is run. The one or more computer-executable components may be at least one software code or a portion thereof. In addition, in this regard, it should be noted that any blocks of the logic flow may represent program steps, or interconnected logic circuits, blocks and functions, or a combination of program steps and logic circuits, blocks and functions. The software may be stored on a physical medium such as a memory chip or memory block implemented within a processor, a magnetic medium such as a hard disk or floppy disk, and an optical medium such as, for example, a DVD and its data variants, a CD, etc. The physical medium is a non-transitory medium.

It should be understood by those skilled in the art that the technical features of the above embodiments may be combined in any manner, and for brevity, all of the possible combinations of the technical features of the above embodiments are not described, however, they should be considered as being within the scope of the description provided herein, as long as there is no contradiction between the combinations of the technical features.

The foregoing examples illustrate only a few embodiments of the invention, which are described in greater detail and are not to be construed as limiting the scope of the invention. It should be noted that it will be apparent to those skilled in the art that various modifications and improvements can be made without departing from the spirit of the invention, which are within the scope of the invention. Accordingly, the scope of the invention should be assessed as that of the appended claims.

Claims (10)

1. A differential InSAR system, comprising:

a laser emitting module for generating and distributing a laser beam;

the radar detection module is used for detecting a target and acquiring and processing signals;

The self-adaptive optical compensation module is used for correcting phase distortion caused by atmospheric turbulence;

The self-adaptive optical compensation module acquires and compensates laser wavefront distortion caused by atmospheric turbulence in real time according to a deep learning technology and a wavefront detection processing mechanism so as to ensure that the differential InSAR system stably and accurately detects in long-distance remote sensing.

2. A differential InSAR system as claimed in claim 1, characterized in that the adaptive optics compensation module comprises an adaptive optics system laser (6-1), a digital micromirror device (6-3) and a wavefront sensor (6-2);

The self-adaptive optical system laser (6-1) emits measuring laser when the tunable laser (1) of the laser emitting module works, after the laser is influenced by atmospheric turbulence, scattered wave fronts are sequentially received by the digital micro-mirror device (6-3) and compression coded, and the wave front sensor (6-2) receives the scattered wave fronts in one exposure time;

Reconstructing continuous images from the one-time exposure result of the wavefront sensor (6-2) by means of the trained deep learning network model so as to improve the frame rate and the closed-loop bandwidth;

and calculating ideal wavefront deviation according to wavefront reconstruction, determining conjugate wavefront execution quantity, and driving a first deformable mirror (2-13) and a second deformable mirror (3-13) of the radar detection module to compensate the radar to receive the laser wavefront so as to achieve the purpose of turbulence resistance.

3. The differential InSAR system of claim 2, wherein when the radar detection module works in a single navigation mode, the calibration light of the adaptive optical compensation module is positioned at a radar transmitting end to detect the phase distortion of the atmospheric turbulence in real time, the wavefront sensor (6-2) and the deformable mirror are arranged in a radar receiving end, the receiving antenna receives the target reflected echo, the wavefront sensor (6-2) firstly measures the distortion calibration light, the compensation coefficient is obtained through an algorithm and sent to the deformable mirror to remove the distortion, and the corrected echo is received by the radar receiving end.

4. A differential InSAR system as defined in claim 1, wherein said laser emitting module includes a tunable laser that splits the laser light into two paths by a beam splitter into said radar detection module.

5. A differential InSAR system according to claim 1, characterized in that the radar detection module comprises a master radar (3) and a slave radar (2);

Two paths of laser emitted by the laser emission module are collimated by a collimator, polarized by a polarizer, aperture filtered and polarized by a half wave plate in respective branches of the master radar (3) and the slave radar (2), the polarized beam splitter divides the local oscillation laser into local oscillation laser and emergent laser, the local oscillation laser enters a receiving light path through a reflector and a quarter wave plate and the polarized state is converted, the emergent laser irradiates a target, reflected light corrects wave front phase through a deformable mirror and then is mixed with the local oscillation laser, an interference pattern and target alignment condition are obtained by observing part of the charge coupled device, and the rest part of the reflected light is used for synthetic aperture processing through an avalanche photodiode detection time sequence after the aperture filtered.

6. A differential InSAR system as claimed in claim 5, characterized in that the slave radar (2) comprises a first collimator (2-1), a first polarizer (2-2), a first aperture (2-3), a first half-wave plate (2-4), a first polarizing beam splitter (2-5), a first quarter-wave plate (2-6), a second quarter-wave plate (2-7), a first mirror (2-8), a second polarizing beam splitter (2-9), a first charge coupled device (2-10), a second aperture (2-11), a first avalanche photodiode (2-12), a first deformable mirror (2-13);

The main radar (3) comprises a second collimator (3-1), a second polarizer (3-2), a third small hole (3-3), a second half-wave plate (3-4), a third polarization beam splitter (3-5), a third quarter-wave plate (3-6), a fourth quarter-wave plate (3-7), a second reflecting mirror (3-8), a fourth polarization beam splitter (3-9), a second charge coupling device (3-10), a fourth small hole (3-11), a second avalanche photodiode (3-12) and a second deformable mirror (3-13).

7. A differential InSAR system according to any of claims 1-6, characterized in that the adaptive optics compensation module employs EFFICIENTSCI deep learning network models based on deep learning techniques.

8. A control method of a differential InSAR system as claimed in any one of claims 1 to 7, comprising:

s00, generating and distributing a laser beam to enter a radar detection module by a laser emission module;

s10, two paths of detection branches of the radar detection module respectively detect targets at the same time and acquire and process echo signals so as to realize a differential detection function;

S20, the self-adaptive optical compensation module acquires and compensates laser wavefront distortion caused by atmospheric turbulence in real time according to a deep learning technology and a wavefront detection processing mechanism.

9. A differential InSAR system as defined in any one of claims 1 to 7 for use in an aircraft.

10. A readable storage medium, characterized in that the readable storage medium has stored therein a computer program comprising program code for controlling a process to execute the process, the process comprising the control method according to claim 8.