Improved hyperbola detection method for extracting tubular cavity characteristics of lava by using moon detection radar

Abstract

The invention provides an improved hyperbola detection method for extracting lava tubular cavity features by using a moon detection radar. The method is characterized in that the lunar detection radar is used for extracting the lava tubular cavity features on the moon. According to the method, firstly, obtained radar original data are preprocessed, and a normal-phase binary matrix and a reverse-phase binary matrix are generated; then, edge enhancement filtering is carried out on the binary matrix through an asymmetric sliding window so as to extract hyperbola features; next, hyperbolic features in the enhanced image are detected. And then the hyperbolic parameters are inverted, and the dielectric constant of the target is calculated and compared with the lunar soil background value. And completing joint verification of cavity features through positive and negative phase hyperbolic space alignment and a phase cancellation mechanism. The application efficiency of the lunar detection radar in the field of geological detection is improved, the problems of low manual processing efficiency, radar image noise interference, experience difference of interpretation personnel and the like are solved, and the accuracy and efficiency of lunar surface lava tube cavity detection are improved.

Description

Technical Field

[0001] The present invention belongs to the field of lunar geological exploration, and more specifically relates to an improved hyperbola detection method for extracting the characteristics of lava tube cavities using a lunar radar. Background Art

[0002] Lava tube cavities widely exist in the subsurface structure of the lunar surface. They are special geological structures formed after magma overflowed during early volcanic activities, and then the surface lava cooled and solidified while the internal lava flow emptied. Under the combined influence of the low lunar gravity environment (about 1 / 6 of that on Earth) and weak weathering, the lateral span of such tube structures can reach hundreds of meters, and the vertical height exceeds ten meters. Their spatial scale is 1 - 3 orders of magnitude larger than that of similar structures on Earth. These natural underground cavities not only preserve the key geological records of lunar volcanic activities, but also, due to their stable thermodynamic environment and good radiation shielding characteristics, are regarded by the international academic community as the preferred candidate sites for building permanent lunar bases. Currently, the detection of lunar surface lava tubes mainly relies on the ground-penetrating radar (GPR) data obtained by the lunar radar (LPR) carried by the Chang'e series of detectors. Existing technical solutions mostly adopt manual interpretation methods, inferring the underground cavity structure by visually identifying the hyperbola diffraction characteristics in radar images. This traditional method has significant defects. The effective detection area of single-track radar data can reach dozens of square kilometers, and the massive data leads to low efficiency in manual processing. At the same time, manual interpretation is also affected by radar image noise interference and differences in the experience of interpreters. Therefore, a complete detection process for the characteristics of lava tubes in radar data is needed to accurately and efficiently detect the distribution of subsurface lava tubes on the moon. Summary of the Invention

[0003] The present invention aims to solve the problem that the current detection of lunar surface lava tubes mainly relies on manually identifying the hyperbola characteristics in lunar radar data. This manual interpretation method has low processing efficiency due to the existence of massive data, and is easily affected by radar image noise interference and differences in the experience of interpreters, thus affecting the accuracy of detection. Therefore, the task of the present invention is to provide an improved hyperbola feature detection method to extract the characteristics of lava tube cavities using a lunar radar and achieve accurate and efficient detection of subsurface lava tubes on the moon.

[0004] To achieve the above object, the present invention is implemented by the following technical solutions: The method includes the following steps:

[0005] Step 1, preprocess the original data obtained by the lunar radar to generate positive and negative binary matrices;

[0006] Step 2, perform edge enhancement filtering on the binary matrix through an asymmetric sliding window;

[0007] Step 3, enhance the hyperbola features in the image through direction detection;

[0008] Step 4, invert the hyperbola parameters and verify the target authenticity based on the dielectric constant;

[0009] Step 5, complete the joint verification of the cavity features through the forward and reverse hyperbola spatial alignment and phase cancellation mechanism.

[0010] In one solution, the preprocessing includes performing positive-phase threshold processing and inverse-phase threshold processing on the radar signal, suppressing the lunar soil noise, and segmenting the bipolar reflection signal.

[0011] In one solution, the spatio-temporal dimensions of the asymmetric sliding window adapt to the hyperbola geometric features, and the isolated noise is suppressed through the extreme point density weight.

[0012] In one solution, the detection of enhancing the hyperbola features in the image improves the detection efficiency by constraining the opening direction of the hyperbola and optimizing the parameter step size.

[0013] In one solution, the dielectric constant verification includes calculating the target dielectric constant and comparing it with the lunar soil background value. If the difference exceeds 20%, it is determined as a cavity.

[0014] In one solution, the phase cancellation mechanism verifies the forward and reverse hyperbolas, judges the amplitude positive and negative, and excludes the crater interference.

[0015] In one solution, the number of sampling points of the asymmetric sliding window on the time axis and the space axis are 3 points and 5 points respectively.

[0016] In one solution, the dielectric constant calculation is realized based on the relationship between the reflection time and the depth.

[0017] Advantages of the present invention:

[0018] 1. Replace manual interpretation through an automated processing flow, improve the processing speed, and be suitable for processing the GB-level data volume of the Chang'e-4 LPR radar.

[0019] 2. Enhance the anti-interference ability, introduce feature enhancement to reduce the hyperbola fitting error, and have an adaptive compensation ability for the signal distortion caused by the lunar soil anisotropy.

[0020] 3. The depth of physical property analysis, combine the dielectric constant inversion results of the forward / reverse hyperbolas, and improve the accuracy of cavity identification. Brief Description of the Drawings

[0021] Figure 1 It is the flow chart of the method of the present invention;

[0022] Figure 2 It is the flow chart of inverting the hyperbola parameters. Detailed implementation manners

[0023] To facilitate the understanding of the present invention, the present invention will be described more comprehensively below with reference to the relevant drawings. Typical embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, these embodiments are provided to make the disclosure of the present invention more thorough and comprehensive.

[0024] Unless otherwise defined, all technical and scientific terms used in the present invention have the same meaning as understood by those skilled in the technical field to which the present invention belongs. The terms used in the description of the present invention in this specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. To facilitate the understanding of the present invention, the present invention will be described more comprehensively below with reference to the relevant drawings. Typical embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, these embodiments are provided to make the disclosure of the present invention more thorough and comprehensive.

[0025] As Figure 1 shown, the specific steps of an improved hyperbola detection method for extracting the characteristics of lava tube cavities using a lunar radar are as follows:

[0026] Step 1, preprocess the original data obtained by the lunar radar to generate positive and negative binary matrices.

[0027] The data processed in this solution is from the lunar radar (LPR) carried by the Chang'e series of lunar probes. The radar emits electromagnetic waves (center frequency 500 MHz) in the form of pulses to the lunar surface, and forms a two-dimensional data matrix I(t, x) after receiving the reflected signals, where: the t-axis: the propagation time of the electromagnetic wave, the sampling interval Δt = 0.3125 ns, the x-axis: the spatial coordinate along the moving direction of the detector, the sampling interval Δx = 0.04 m, and the matrix element I(t i , x j ): represents the signal amplitude received at position t i , time x j .

[0028] Threshold segmentation is used to perform positive / negative bipolar processing on the original matrix:

[0029] Positive-phase threshold processing:

[0030]

[0031] Negative-phase threshold processing:

[0032]

[0033] Where: th = 0.8: Threshold coefficient to suppress the random noise on the lunar soil surface. Output the positive / negative binarized matrix I p and I n , which is characterized by: effectively retaining strong reflection signals (such as the reflection of the lava tube roof) with amplitudes exceeding the threshold. Eliminating noise signals (such as lunar soil particle scattering) with amplitudes lower than the threshold.

[0034] Step 2, perform edge enhancement filtering on the binarized matrix through an asymmetric sliding window; adopt asymmetric sliding window average filtering (window size 3×5), and enhance the continuity of the hyperbola edge through density weight assignment to suppress isolated noise points. The specific process is as follows:

[0035] S201. Filter kernel definition, generate a 3×5 rectangular mean filter kernel:

[0036]

[0037] Window size description: 3 rows: Cover 3 sampling points (about 0.94 ns) along the time axis (signal propagation direction), matching the time span of the hyperbola vertex. 5 columns: Cover 5 sampling points (about 0.2 m) along the space axis (the moving direction of the rover), matching the horizontal extension range of the hyperbola.

[0038] S202. Density weight calculation, for each position (i,j), calculate the density weight of the extreme points in its 3×5 neighborhood:

[0039]

[0040] For extreme points, still retain the weight of 1. If the extreme points in the neighborhood are dense (such as the hyperbola edge), then I be (i,j) → 1, which can be used as a springboard for local broken hyperbolas to continue searching for hyperbola points. If the extreme points in the neighborhood are isolated (such as noise), then I be (i,j) → 0 (for example, when only 1 point is 1, I be = 0.067)

[0041] S203. Processing effect, retaining dense areas: The hyperbola edge has a high density weight due to the continuous distribution of extreme points. Suppressing isolated noise: Random noise points have a low weight due to the low density of the neighborhood.

[0042] Step 3, detect the hyperbola features in the enhanced image;

[0043] S301. Platform point detection

[0044] Based on the edge-enhanced binary matrix I be, search for the platform points corresponding to the vertices of the hyperbola along the time axis (matrix row direction): Definition of platform points: A set of extreme points continuously distributed horizontally, serving as the candidate area for the top of the hyperbola.

[0045] Detection process:

[0046] (1) Scan row by row: Scan each row of data from left to right, and mark the area of continuous extreme points (i.e., the continuous columns where I be (i, j) = 1).

[0047] (2) Platform trigger condition: If a complete continuous extreme point area is detected, trigger the subsequent search process for the inclined branch.

[0048] (3) Dynamic platform length: The platform length is determined by the actual number of continuous extreme points, without a fixed threshold limit, but it needs to meet the requirements for the number of subsequent inclined points.

[0049] S302. Search for the inclined points of the hyperbola. For each detected platform point area, search for the inclined branches of the hyperbola to the left and right:

[0050] Definition of key parameters:

[0051]

[0052]

[0053] Branch search algorithm:

[0054] Independent search for left and right branches:

[0055] Search for the left inclined branch to the left from the left end point of the platform (direction = -1)

[0056] Search for the right inclined branch to the right from the right end point of the platform (direction = 1)

[0057] Slope tracking and fault tolerance mechanism:

[0058] Main path tracking: Search for extreme points row by row downward along the time axis (the direction of increasing row number), and preferentially select adjacent points (j + direction) that are consistent with the current direction.

[0059] Fuzzy fault tolerance: If there are no valid points in the main direction, allow lateral offset within the horizontal fuzzy tolerance (H blur ) range, and skip at most H count points.

[0060] Depth verification: The time span of a single-sided branch needs to satisfy Δt ≥ d min (i.e., the number of rows ≥ depth).

[0061] Inclination point counting: Accumulate the number of valid inclination points in the left and right branches. If the total number N left +N right ≥N min (sum = 50), then retain this hyperbola candidate.

[0062] As Figure 2 shown, in step 4, invert the hyperbola parameters and verify the target authenticity based on the dielectric constant;

[0063] S401. Coordinate transformation: Convert the pixel coordinates (i, j) to physical coordinates (t, x):

[0064] t = i·Δt, x = j·Δx

[0065] where Δt = 0.3125 ns and Δx = 0.04 m (code variables row_spacing, col_spacing).

[0066] S402. Hyperbola mathematical model

[0067] The reflected signal from the top plate of the void pipe appears as a hyperbola in the radar image, and its equation can be expressed as:

[0068]

[0069] where: x 0 : Horizontal position of the hyperbola vertex (corresponding to the center of the platform), h: Burial depth of the top plate of the void pipe, v: Propagation speed of electromagnetic waves in lunar soil ( c = 0.3 m / ns is the speed of light in vacuum, ε is the dielectric constant of lunar soil)

[0070] S403. Weighted least squares fitting: Perform nonlinear fitting on the detected hyperbola point set {(x k , t k )}:

[0071] S4031. Weight assignment: Multiple observation points at the same time t share the weight, and the weight value is w k = 1 / N t , where N t is the number of points at time t k . Reduce the excessive influence of densely sampled points on fitting and enhance the weight of the horizontal distribution

[0072] S4032. Fitting process:

[0073] Define the fitting model:

[0074]

[0075] Initial parameter estimation:

[0076] v (0) = c,

[0077] Objective function:

[0078]

[0079] S4033. Verification of fitting results:

[0080] Coefficient of determination R 2 ≥ 0.7 (filtering low-quality fitting).

[0081] Dielectric constant calculation:

[0082]

[0083] where c 0 = 0.3 m / ns is the speed of light in vacuum. When the difference between ε r and the lunar soil background value (typical value 3 - 5) exceeds 20%, it is determined as a cavity target.

[0084] Step 5. Complete the joint verification of cavity features through the forward and reverse hyperbola space alignment and phase cancellation mechanism.

[0085] Forward and reverse hyperbola matching: Perform space alignment verification on the detected forward hyperbola H + and reverse hyperbola H - to effectively distinguish real cavities from interference targets such as meteorite impact craters through the phase cancellation mechanism. Output the successfully matched hyperbola pairs. Forward hyperbola: Reflection signal of the lava tube roof. Reverse hyperbola: Reflection signal of the lava tube floor.

[0086] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The described program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM), etc.

[0087] It should be understood that the detailed description of the technical solutions of the present invention with the aid of the preferred embodiments above is illustrative rather than restrictive. Those of ordinary skill in the art can modify the technical solutions recorded in each embodiment on the basis of reading the specification of the present invention, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of each embodiment of the present invention.

Claims

1. An improved hyperbola detection method for extracting lava tube features using lunar radar, characterized in that: The method comprises the following steps: Step 1: preprocess the raw data obtained by the lunar radar to generate positive phase and negative phase binary matrices; Step 2, edge enhancement filtering is performed on the binarized matrix through an asymmetric sliding window; Step 3, detecting hyperbolic features in the enhanced image; Step 4, invert the hyperbolic parameters and verify the target authenticity based on the dielectric constant; Step 5: Complete the joint verification of hole features through the positive and negative hyperbolic space alignment and phase cancellation mechanism.

2. The improved hyperbola detection method for extracting lava tube features using lunar radar according to claim 1, characterized in that: The preprocessing includes performing positive phase threshold processing and negative phase threshold processing on the radar signal, suppressing lunar soil noise and segmenting out bipolar reflection signals.

3. The improved hyperbola detection method for extracting lava tube features using lunar radar according to claim 1, characterized in that: The space-time dimension of the asymmetric sliding window is adapted to the hyperbolic geometric features, and isolated noise is suppressed by the extreme point density weight.

4. The improved hyperbola detection method for extracting lava tube features using lunar radar according to claim 1, characterized in that: The hyperbolic features in the detection enhancement image improve detection efficiency by constraining the hyperbolic opening direction and optimizing the parameter step size.

5. The improved hyperbola detection method for extracting lava tube features using lunar radar according to claim 1, characterized in that: The dielectric constant verification includes calculating the target dielectric constant and comparing it with the lunar soil background value. If the difference exceeds 20%, it is judged to be a cavity.

6. The improved hyperbola detection method for extracting lava tube features using lunar radar according to claim 1, characterized in that: The phase cancellation mechanism verifies the positive and negative phase hyperbolas, determines the positive and negative amplitudes, and eliminates crater interference.

7. The improved hyperbola detection method for extracting lava tube features using lunar radar according to claim 3, characterized in that: The asymmetric sliding window has 3 sampling points and 5 sampling points on the time axis and space axis respectively.

8. The improved hyperbola detection method for extracting lava tube features using lunar radar according to claim 5, characterized in that: The dielectric constant calculation is implemented based on the relationship between reflection time and depth.

Images