Submarine topography photon extraction method and device, storage medium and product

Abstract

The invention discloses a submarine topography photon extraction method and device, a storage medium and a product. The method comprises the steps that sea surface photon data and water body photon data are extracted from shallow sea segment photon data; initializing a major semi-axis and a minor semi-axis of the elliptic filtering kernel; calculating a first density value of each photon in the water body photon data; determining a first density threshold value according to the initial values of the major semi-axis and the minor semi-axis; performing primary filtering on the water body photon data; updating a major semi-axis and a minor semi-axis according to the elevation of each photon; calculating a second density value of photons in the water body photon data after primary filtering; calculating a second density threshold value of the second elevation segment according to the first density value; calculating a third density threshold according to the updated semi-major axis; and performing secondary filtering according to the second density threshold, the third density threshold and the second density value to obtain submarine topography photon data. According to the invention, the filtering accuracy is improved, and the problems of missing detection and false detection are improved.

Description

Technical Field

[0001] The present invention belongs to the field of satellite-borne laser radar photon point cloud processing, and in particular relates to a method, device, storage medium and product for extracting seabed terrain photons taking spatial relationships into consideration. Background Art

[0002] Since the ATLAS (Advanced Topographic Laser Altimeter System) carried by the ICESat-2 satellite uses micro-pulse photon technology, it is easily affected by factors such as the solar background and atmospheric scattering, resulting in the received echo signals containing varying degrees of noise photons. Therefore, in order to obtain water depth data from the ATLAS data, it is critical to filter the original ATLAS data and extract the seabed topography photons.

[0003] The study found that it is more practical to use a local parameter estimation method to extract photons from shallow sea terrain. Reference 1 (see: Zhu Xiaoxiao, Nie Sheng, Wang Cheng, Xi Xiaohuan, Hu Zhenyue. A Ground Elevation and Vegetation Height Retrieval Algorithm Using Micro-Pulse Photon-Counting Lidar Data[J]. Remote Sensing, 2018, 10(12).) takes slope adaptation into account in the calculation of photon density, and uses Gaussian fitting to calculate the photon density threshold of the photon density histogram to extract signal photons. This method does not take into account the complex distribution of signal photons, and has poor results for extracting sparsely distributed signal photons. Reference 2 (see: Chen Yifu, Le Yuan, Zhang Dongfang, Wang Yong, Qiu Zhenge, Wang Lizhe. A photon-counting LiDAR bathymetric method based on adaptive variable ellipse filtering [J]. Remote Sensing of Environment, 2021, 256.) utilizes the density difference between signal photons and noise photons in the ATL03 data, and adaptively adjusts the size of the elliptical filter kernel and the density threshold according to the photon depth, thereby realizing the recognition and extraction of photons on the sea, sea surface and seabed surface. The application effect on high signal-to-noise ratio photon data is very good, but the relationship between photon density and water depth is simplified to linear, which weakens the generalization ability of the model and will introduce more noise when applied to low signal-to-noise ratio data.Reference 3 (Hsu, H., Huang, C., Jasinski, MF, Li, Y., Gao, H., Yamanokuchi, T., Wang, C., Chang, TM, Ren, H., Kuo, C., & Tseng, KA semi-empirical scheme for bathymetric mapping in shallow water by ICESat-2and Sentinel-2: A case study in the South China Sea. ISPRS Journal of Photogrammetry and Remote Sensing, 178, 2021, 1-19.) utilizes the difference in the aggregation of signal photons and noise photons in ATL03 data, and adopts a bimodal Gaussian fitting algorithm to identify and extract sea surface and seabed photons in the windows divided along the track. The extraction effect is mainly affected by the sharpness of the peak in the window. For low signal-to-noise ratio data with weak aggregation of sea surface and seabed signal photons, the fitting effect is poor and a lot of noise will be introduced. Reference 4 (see: N.Xu, Y.Ma, H.Zhou, W.Zhang, Z.Zhang and XHWang,"A Method to DeriveBathymetry for Dynamic Water Bodies Using ICESat-2and GSWD Data Sets,"in IEEE Geoscience and Remote Sensing Letters, vol.19, pp.1-5, 2022, Art no.1500305, doi:10.1109 / LGRS.2020.3019396.) uses an adaptive DBSCAN algorithm (Density-Based Spatial Clustering of Applications with Noise) to extract underwater photons from ICESat-2 data. This algorithm may misjudge local areas where photon data is sparse and the terrain changes dramatically.Reference 5 (see: P. Yang, H. Fu, J. Zhu, Y. Li and C. Wang, "An Elliptical Distance Based Photon Point Cloud Filtering Method in Forest Area," in IEEE Geoscience and Remote Sensing Letters, vol. 19, pp. 1-5, 2022, Art no. 6504705, doi: 10.1109 / LGRS.2021.3124612.) uses a slope-adaptive fixed ellipse to traverse the photons to calculate the photon neighborhood distance and then identifies and extracts signal photons based on the distance threshold, taking into account the geographical similarity between adjacent photons, but the application effect is not good for the extraction of discretely distributed terrain photons. Reference 6 (see: ZHANG Guoping, XU Qing, XING Shuai, et al. A noise-removal algorithm without input parameters based on quadtree isolation for photon-counting LiDAR [J]. IEEE Geoscience and Remote Sensing Letters, 2021, 19: 1-5.) uses the MATLAS data simulated from ICESat-2 to perform quadtree denoising without input parameters to complete the recognition and extraction of point cloud signals. This method has a high misrecognition rate for noise photons under strong noise background. In response to the problems in reference 6, reference 7 (see: Liu Xiang, Zhang Lihua, Dai Zeyuan, et al. A method for ICESat-2 point cloud denoising under strong noise background without input parameters [J]. Acta Photonica Sinica, 2022, 51(11): 346-356.) uses ATL03 data, performs pruning on the basis of quadtree establishment, and adds secondary denoising to improve the extraction effect of signal photons.Reference 8 (see: G. Zhang, W. Lian, S. Li, H. Cui, M. Jing and Z. Chen, "A Self-Adaptive Denoising Algorithm Based on Genetic Algorithm for Photon-Counting Lidar Data," in IEEE Geoscience and Remote Sensing Letters, vol. 19, pp. 1-5, 2022, Art no. 6501405, doi: 10.1109 / LGRS.2021.3067609.) calculates the neighbor distance based on the weighted distance of the photon neighbors and extracts the signal photons as features. This method is unstable when applied to photon data with low signal-to-noise ratio.

[0004] In the method of extracting photons from seabed topography, the key is the selection of the shape and size of the density statistical unit and the setting of the density segmentation threshold. In recent years, the shape of the filter kernel has evolved from circle to ellipse to parallelogram, and its size and density segmentation threshold have evolved from fixed to adaptive. However, most methods simplify the calculation of photon density by establishing a linear relationship with water depth, and rarely consider the geographical differences of photon data. As a result, these methods have good results in extracting seabed topography for photon data with high signal-to-noise ratio, but the processing results for photon data with low signal-to-noise ratio will result in more missed terrain photons and more falsely detected noise photons. Summary of the invention

[0005] The purpose of the present invention is to provide a method, device, storage medium and product for extracting photons from seabed topography, so as to solve the problem that traditional methods do not take into account the geographical differences of photon data, resulting in missed detection and false detection of photon data with low signal-to-noise ratio.

[0006] The present invention solves the above technical problems through the following technical solutions: a method for extracting photons from seabed topography, comprising:

[0007] Preprocess ICESat-2 / ATL03 photon data to obtain shallow sea photon data;

[0008] Extracting sea surface photon data and water body photon data from the shallow sea section photon data;

[0009] A mathematical model of an elliptical filter kernel is constructed according to the along-track distance and elevation of the photons, and the major semi-axis and the minor semi-axis of the elliptical filter kernel are initialized;

[0010] Calculate a first density value of each photon in the water body photon data according to the initial values ​​of the major semi-axis and the minor semi-axis and a mathematical model of an elliptical filter kernel;

[0011] Determining a first density threshold according to the water body photon data and initial values ​​of the major semi-axis and the minor semi-axis;

[0012] Filtering the water body photon data once according to the first density threshold and the first density value of each photon;

[0013] The major semi-axis and the minor semi-axis are updated according to the elevation of each photon in the water body photon data after the primary filtering; the second density value of the corresponding photon in the water body photon data after the primary filtering is calculated according to the updated major semi-axis and minor semi-axis and the mathematical model of the elliptical filter kernel;

[0014] Calculating a second density threshold for each second elevation segment based on the first density value of each photon;

[0015] Calculate the third density threshold based on the updated major semi-axis;

[0016] The water body photon data after the primary filtering is subjected to secondary filtering according to the second density threshold, the third density threshold and the second density value of each photon to obtain seabed topography photon data.

[0017] Further, extracting sea surface photon data and water body photon data from the shallow sea section photon data includes:

[0018] The width of the first elevation segment is set, the photon data of the shallow sea segment is projected according to the elevation, and the number of photons in each first elevation segment is counted to form an elevation distribution histogram;

[0019] The elevation distribution histogram is fitted using a bimodal Gaussian distribution fitting algorithm, and then the sea surface photon data and the water body photon data are extracted.

[0020] Further, the major semi-axis and the minor semi-axis of the elliptical filter kernel are initialized according to the sea surface photon data, including:

[0021] Dividing the sea surface photon data into n segments along the track according to the step length, and calculating the maximum elevation difference of each first segment along the track;

[0022] Calculate the initial values ​​of the major and minor axes of the elliptical filter kernel. The specific formula is:

[0023] b0=|H max -H min |;

[0024] a0=R ab b0,

[0025] Among them, b0 represents the initial value of the minor semi-axis of the elliptical filter kernel, H max represents the upper boundary of the sea surface photon data, H minrepresents the lower boundary of the sea surface photon data, a0 represents the initial value of the major semi-axis of the elliptical filter kernel, R ab represents the shape adjustment factor of the elliptical filter kernel, Δd represents the step size, ΔSH i Indicates the maximum elevation difference of the i-th first segment along the track.

[0026] Furthermore, the specific process of calculating the first density value or the second density value of each photon includes:

[0027] Calculate the distance between each photon and other photons according to the mathematical model of the elliptical filter kernel;

[0028] A first distance threshold is set; for each photon, the number of photons whose distance does not exceed the first distance threshold is counted, and the number of photons is the first density value or the second density value of the photon.

[0029] Furthermore, the formula for determining the first density threshold is:

[0030]

[0031]

[0032] Among them, D T1 represents the first density threshold, SN1 represents the signal photon distribution density within the initial unit elliptical kernel area, SN2 represents the noise photon distribution density within the initial unit elliptical kernel area, a0 represents the initial value of the major semi-axis of the elliptical filter kernel, b0 represents the initial value of the minor semi-axis of the elliptical filter kernel, M represents the number of photons in the water body photon data, ΔH represents the vertical distance between the highest photon and the lowest photon in the water body photon data, ΔL represents the difference between the maximum along-track distance and the minimum along-track distance in the water body photon data, m l Represents the number of photons within the elevation range of the noise photon distribution in the water body photon data, N H Indicates the elevation range of the noise photon distribution in the set water body photon data.

[0033] Furthermore, the updating formulas of the major semi-axis and the minor semi-axis are:

[0034] b=b0+τ|h p -H min |, a=R ab b;

[0035] Among them, b represents the updated minor semi-axis, b0 represents the initial value of the minor semi-axis of the ellipse filter kernel, τ represents the scaling factor of the ellipse, and h p represents the height of the pth photon, H min represents the lower boundary of the sea surface photon data, a represents the updated major semi-axis, and R ab Represents the shape adjustment factor of the elliptical filter kernel.

[0036] Further, calculating a second density threshold in each second elevation segment according to the first density value of each photon includes:

[0037] The width of the second elevation segment is set, the water body photon data after the primary filtering is projected according to the elevation, the first density values ​​of all photons in each second elevation segment are counted, and the first density values ​​are arranged in descending order;

[0038] The maximum inter-class variance method is used to calculate the initial density threshold of each second elevation segment;

[0039] The second density threshold of the corresponding second elevation segment is calculated according to the initial density threshold of each second elevation segment and the initial density thresholds of its adjacent second elevation segments.

[0040] Furthermore, the extraction method also includes using a bidirectional seed point growth algorithm to supplement the seabed topography photon data, specifically including:

[0041] Calculating the difference in along-track distances between every two adjacent photons in the seabed topography photon data;

[0042] Calculate the second distance threshold according to the difference between the along-track distances of every two adjacent photons;

[0043] Determine the discontinuous area according to the difference in along-track distances between every two adjacent photons and a second distance threshold;

[0044] For discontinuous areas, one of the two adjacent photons is used as the initial seed point, and the nearest neighbor photon search is performed in the direction of the other photon to obtain the candidate photon;

[0045] Determine whether the candidate photon is a signal photon or a noise photon; when the candidate photon is a signal photon, add the signal photon to the seabed topography photon data, and use the signal photon as a new seed point, repeat the steps of nearest neighbor photon search and candidate photon judgment until the seed point exceeds the discontinuous area; when the candidate photon is a noise photon, complete the processing of one side of the discontinuous area;

[0046] Taking the other photon of the two adjacent photons as the initial seed point, a nearest neighbor photon search is performed in the direction of one of the photons to obtain a candidate photon;

[0047] Determine whether the candidate photon is a signal photon or a noise photon; when the candidate photon is a signal photon, add the signal photon to the seabed topography photon data, and use the signal photon as a new seed point to repeat the steps of nearest neighbor photon search and candidate photon judgment until the candidate photon exceeds the discontinuous area; when the candidate photon is a noise photon, complete the processing of the other side of the discontinuous area;

[0048] Merging the seabed topography photon data obtained by processing both sides of the discontinuous area, and dividing the merged seabed topography photon data into multiple segments along the track to obtain multiple second along-track segments;

[0049] For each of the second along-track segments, multiple third elevation segments are obtained according to the elevation segmentation, the number of photons in each third elevation segment is counted, and the third elevation segment with the largest number of photons is found;

[0050] Determine whether the number of photons in the third elevation segment with the largest number of photons is greater than twice the photon data of its two adjacent third elevation segments. If so, retain the third elevation segment with the largest number of photons and discard the two adjacent third elevation segments of the third elevation segment with the largest number of photons; if not, reduce the width of the third elevation segment, and repeat the steps of segmenting by elevation and determining until a third elevation segment with a number of photons greater than twice the photon data of its two adjacent third elevation segments is found;

[0051] All the processed second along-track segments are merged to obtain the final seabed topography photons.

[0052] Based on the same concept, the present invention also provides an electronic device, including a memory, a processor, and a computer program / instruction stored in the memory, wherein the processor executes the computer program / instruction to implement the above-mentioned method for extracting photons from seabed topography.

[0053] Based on the same concept, the present invention also provides a computer-readable storage medium on which a computer program / instruction is stored. When the computer program / instruction is executed by a processor, the above-mentioned method for extracting photons from seabed topography is implemented.

[0054] Based on the same concept, the present invention also provides a computer program product, including a computer program / instruction, which implements the above-mentioned method for extracting photons from seabed topography when executed by a processor.

[0055] Beneficial Effects

[0056] Compared with the prior art, the advantages of the present invention are:

[0057] The present invention first calculates a first density threshold according to the initial values ​​of the major and minor semi-axis of the elliptical filter kernel, and the first density threshold filters the water body photon data once to achieve global filtering of the water body photon data; then, considering the variability and diversity of the spatial distribution of photons, the density threshold (i.e., the second density threshold) of each elevation segment is calculated, and a nonlinear relationship between the photon density and the water depth is established, which can better reflect the density difference between the signal photons and the noise photons in each elevation segment, improve the accuracy of filtering, and improve the problems of missed detection and false detection.

[0058] Taking into account that each elevation segment may have signal photons and noise photons of comparable density, a bidirectional seed point growth algorithm is used to supplement the seabed topography photon data, that is, the signal photons filtered out by the secondary filtering are added to the seabed topography photon data, which further improves the accuracy of the filtering and avoids missed detection and false detection problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solution of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only one embodiment of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0060] Figure 1 is a flow chart of a method for extracting photons from seabed topography in an embodiment of the present invention;

[0061] Figure 2 is an elevation distribution histogram in an embodiment of the present invention;

[0062] Figure 3 is a schematic diagram of sea surface photon data in an embodiment of the present invention; wherein the two straight lines are the upper and lower boundary lines of the sea surface photon data respectively;

[0063] Figure 4 is a schematic diagram of a filtering result in an embodiment of the present invention;

[0064] Figure 5 is a schematic diagram of elliptical filter kernels for different photons in an embodiment of the present invention;

[0065] Figure 6 In the embodiment of the present invention, the water body photon data after the primary filtering is segmented in elevation;

[0066] Figure 7 is the second density threshold in each second elevation segment (or vertical water depth segment) in the embodiment of the present invention;

[0067] Figure 8 is a schematic diagram of secondary filtering results in an embodiment of the present invention; wherein the green dots represent seabed topography photons obtained by secondary filtering;

[0068] Fig. 9 : is a schematic diagram of discontinuous region positioning in an embodiment of the present invention; wherein each rectangular box is a discontinuous region, and the blue point and the red point in the rectangular box are two adjacent photons that determine the discontinuous region;

[0069] Fig.10 It is the supplementary result of leftward seed point growth in the embodiment of the present invention;

[0070] Fig.11It is the rightward seed point growth supplement result in the embodiment of the present invention;

[0071] Fig.12 It is the supplementary result of the growth of the seed points after the left and right directions are merged in the embodiment of the present invention;

[0072] Fig.13 Schematic diagram of the second along-track segmentation and the third elevation segmentation in an embodiment of the present invention;

[0073] Fig.14 is the final seabed topography photon data obtained after coarse denoising in the embodiment of the present invention;

[0074] Fig.15 is the extraction result of the first ICESat-2 / ATL03 photon data by the present invention in an embodiment of the present invention;

[0075] Fig.16 is the extraction result of the second ICESat-2 / ATL03 photon data by the present invention in an embodiment of the present invention;

[0076] Fig.17 is the extraction result of the third ICESat-2 / ATL03 photon data in the embodiment of the present invention;

[0077] Fig.18 It is the extraction result of the 4th ICESat-2 / ATL03 photon data by the present invention in the embodiment of the present invention. DETAILED DESCRIPTION

[0078] The following is a clear and complete description of the technical solutions in the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0079] The technical solution of the present application is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0080] Example 1

[0081] like Figure 1 As shown, a method for extracting photons from seabed topography provided by an embodiment of the present invention comprises the following steps:

[0082] Step S1: Preprocess the ICESat-2 / ATL03 photon data to obtain shallow sea segment photon data.

[0083] Download ICESat-2 / ATL03 photon data from the public NSIDC (National Snow and Ice Data Center, U.S. National Snow and Ice Data Center), ICESat-2 / ATL03 photon data is HDF5 (Hierarchical Data Format) file. Read the key photon information (such as longitude and latitude, photon height (or elevation), and along-track distance) in the HDF5 file to form a multidimensional list, and then perform preliminary data cleaning according to longitude and latitude and elevation, remove photon data that are not in the target area, non-sea area data with too high or too low elevation, and noise data; then use photon elevation and along-track distance as coordinate axes for visualization, visually determine the area with seabed reflection signals, and cut according to the along-track distance range of the corresponding area to complete photon data cleaning and obtain shallow sea section photon data.

[0084] Step S2: extracting sea surface photon data and water body photon data from the shallow sea section photon data.

[0085] In a specific embodiment of the present invention, extracting sea surface photon data and water body photon data from shallow sea section photon data includes:

[0086] Step S2.1: Set the width of the first elevation segment, project the shallow sea segment photon data according to the elevation, count the number of photons in each first elevation segment, and form an elevation distribution histogram, such as Figure 2 As shown;

[0087] Step S2.2: Use a bimodal Gaussian distribution fitting algorithm to fit the elevation distribution histogram, and then extract the sea surface photon data and water body photon data.

[0088] In this embodiment, the width of the first elevation segment is 0.1m. The size of each first elevation segment is:

[0089]

[0090] Among them, V g Indicates the size of the g-th first elevation segment, H 1g Indicates the elevation range of the g-th first elevation segment, n g represents the number of photons in the g-th first elevation segment. The expression of the double-peaked Gaussian function is:

[0091]

[0092] Among them, fg(h) represents a double-peak Gaussian function; h represents the photon elevation; A1 and A2 represent the amplitudes of the double peaks, which are approximately the maximum number of sea surface photons and the maximum number of seabed photons that fall within the first elevation segment; μ1 and μ2 represent the expectations of the double peaks (μ2<μ1), that is, the approximate sea surface elevation and seabed elevation; σ1 and σ2 represent the standard deviations of the double peaks, that is, the standard deviation of the sea surface photon peak and the standard deviation of the seabed photon peak, respectively.

[0093] The photon noise above the sea surface is: h>μ1+2σ1;

[0094] The photon data of the sea surface is: μ1-2σ1≤h≤μ1+2σ1;

[0095] The photon data of water body is: h<μ1-2σ1;

[0096] The upper and lower boundaries of sea surface photons are: H max =μ1+2σ1,H min =μ1-2σ1, where H max represents the upper boundary of the sea surface photon data, H min Represents the lower boundary of the sea surface photon data.

[0097] The standard deviation σ2 is used to characterize the signal-to-noise ratio of photon data: when σ2≤1, the signal-to-noise ratio is high, and when σ2>1, the signal-to-noise ratio is low.

[0098] After the sea surface photons are extracted, the remaining water body photons also include seabed photons and noise, such as Figure 3 shown.

[0099] Step S3: construct a mathematical model of the elliptical filter kernel according to the along-track distance and elevation of the photons, and initialize the major semi-axis and minor semi-axis of the elliptical filter kernel.

[0100] In this embodiment, the mathematical model of the elliptical filter kernel is expressed as:

[0101]

[0102] Among them, dist(p,q) represents the distance between photon p and photon q, a represents the major semi-axis of the elliptical filter kernel, b represents the minor semi-axis of the elliptical filter kernel, and d p d q represent the distances along the track of photon p and photon q, respectively, and h p 、h q They represent the elevations of photon p and photon q respectively.

[0103] There are two ways to initialize the major and minor semi-axis of the elliptical filter kernel. The first is to set the initial values ​​of the major and minor semi-axis of the elliptical filter kernel based on experience, and the second is to calculate the initial values ​​of the major and minor semi-axis of the elliptical filter kernel based on sea surface photon data. Using the first method, when processing different ICESat-2 / ATL03 photon data, it is necessary to manually set appropriate empirical values ​​each time, which cannot be widely used. Using the second method, the initial values ​​of the major and minor semi-axis can be adaptively calculated based on the extracted sea surface photon data, without the need for empirical value setting, and has a wide range of applications. It also takes into account spatial correlation and has higher accuracy. Therefore, an embodiment of the present invention calculates the initial values ​​of the major and minor semi-axis of the elliptical filter kernel based on sea surface photon data, specifically including:

[0104] Step S3.1: Divide the sea surface photon data into n segments along the track according to the step length Δd, and calculate the maximum elevation difference ΔSH of each first segment along the track i , that is, the difference between the maximum elevation and the minimum elevation in each first track segment.

[0105] Step S3.2: Calculate the initial values ​​of the major and minor axes of the elliptical filter kernel. The specific formula is:

[0106] b0=|H max -H min | (4)

[0107]

[0108] Among them, b0 represents the initial value of the minor semi-axis of the elliptical filter kernel, H max represents the upper boundary of the sea surface photon data, H min represents the lower boundary of the sea surface photon data, a0 represents the initial value of the major semi-axis of the elliptical filter kernel, R ab represents the shape adjustment factor of the elliptical filter kernel, Δd represents the step size, ΔSH i represents the maximum elevation difference of the first track segment of the ith track. In this embodiment, the step length Δd is 7m.

[0109] Step S4: Calculate the first density value of each photon in the water body photon data according to the initial values ​​of the major semi-axis and the minor semi-axis and the mathematical model of the elliptical filter kernel.

[0110] First, the initial values ​​of the major and minor semi-axis calculated by formula (4) and formula (5) are substituted into formula (3), and then the distance dist(p, q) between each photon p and other photons q in the water body photon data is calculated, and the number of photons whose distance dist(p, q) does not exceed the first distance threshold is counted. Then, the first density value of photon p is the number of photons whose distance dist(p, q) does not exceed the first distance threshold. In this embodiment, the first distance threshold is 1m.

[0111] Step S5: Determine a first density threshold according to the water body photon data and the initial values ​​of the major semi-axis and the minor semi-axis.

[0112] In this embodiment, the formula for determining the first density threshold is:

[0113]

[0114]

[0115] Among them, D T1 represents the first density threshold, SN1 represents the signal photon distribution density within the initial unit elliptical core area, SN2 represents the noise photon distribution density within the initial unit elliptical core area, M represents the number of photons in the water body photon data, ΔH represents the vertical distance between the highest photon and the lowest photon in the water body photon data, ΔL represents the difference between the along-track distance of the leftmost photon and the along-track distance of the rightmost photon in the water body photon data (i.e., the difference between the maximum along-track distance and the minimum along-track distance), m l Represents the number of photons within the elevation range of the noise photon distribution in the water body photon data, N H Indicates the elevation range of the noise photon distribution in the set water body photon data. In this embodiment, N H Set to 5m.

[0116] Step S6: filtering the water body photon data according to the first density threshold and the first density value of each photon.

[0117] Photons with a first density value less than the first density threshold are filtered out to achieve global filtering of water body photon data, such as Figure 4 shown. Figure 4 In the figure, the ellipse is the elliptical filter kernel, the blue dots are the noise photons filtered out by the first filter, the brown dots are the water photons retained after the first filter, and the two straight lines represent the elevation range N of the noise photon distribution. H .

[0118] Step S7: Update the major semi-axis and the minor semi-axis according to the elevation of each photon in the water body photon data after the first filtering; calculate the second density value of the corresponding photon in the water body photon data after the first filtering according to the updated major semi-axis and minor semi-axis and the mathematical model of the elliptical filter kernel.

[0119] For the photon p in the water body photon data after one filtering, the update formulas of the major semi-axis and the minor semi-axis are:

[0120] b=b0+τ|h p -H min |, a=R ab b (8)

[0121] Where b represents the updated minor semi-axis, b0 represents the initial value of the minor semi-axis of the ellipse filter kernel, τ represents the scaling factor of the ellipse (τ>0), and h p represents the elevation of the pth photon in the water body photon data after one filtering, H min represents the lower boundary of the sea surface photon data, a represents the updated major semi-axis, and R ab Represents the shape adjustment factor of the elliptical filter kernel. R ab >1, shape adjustment factor R ab It can be set based on experience or calculated according to formula (5).

[0122] Substitute the major semi-axis a and the minor semi-axis b calculated by formula (8) into formula (3) to obtain the elliptical filter kernel centered on photon p, and then calculate the second density value of photon p: use the elliptical filter kernel centered on photon p to calculate the distance dist(p, q) between photon p and other photons q in the water body photon data after one filtering, and count the number of photons whose distance dist(p, q) does not exceed the first distance threshold. Then the second density value of photon p is the number of photons whose distance dist(p, q) does not exceed the first distance threshold. Figure 5 The elliptical filter kernel corresponding to different photons is shown. Figure 5 It can be seen that the elliptical filter kernel increases with the increase of water depth (ie, the difference between the elevation of the photon and the lower boundary of the sea surface photon data).

[0123] The present invention adaptively updates the major semi-axis and the minor semi-axis according to formula (8), and then generates an adaptive elliptical filter kernel, and uses the adaptive elliptical filter kernel to calculate the density value of the photon (i.e., the second density value), taking into account the complex and unknown nonlinear relationship between the photon density and the water depth that cannot be directly expressed by a formula, which is beneficial to the filtering processing of photon data with low signal-to-noise ratio.

[0124] Step S8: Calculate the second density threshold of each second elevation segment according to the first density value of each photon.

[0125] In a specific embodiment of the present invention, calculating the second density threshold in each second elevation segment according to the first density value of each photon includes:

[0126] Step S8.1: Set the width of the second elevation segment, project the filtered water body photon data according to the elevation, count the first density values ​​of all photons in each second elevation segment, and arrange the first density values ​​in descending order.

[0127] Figure 6 Each second elevation segment is shown, each rectangle is a second elevation segment, and the width of each second elevation segment is 1m; Figure 7 Each second elevation segment S is shown kThe median of the first density value and the second density threshold, where the green line represents the corresponding second elevation segment S k The red line indicates the median of the first density value in the second elevation segment S. k The second density threshold of

[0128] Step S8.2: Calculate each second elevation segment S using the maximum inter-class variance method (OTSU algorithm) k The initial density threshold includes:

[0129] Calculate the inter-class variance between "signal" photons and "noise" photons. The specific formula is:

[0130]

[0131]

[0132]

[0133]

[0134] μ(t)=w s (t)×μ s (t)+w n (t)×μ n (t) (13)

[0135] (σ(t)) 2 =w s (t)×(μ s (t)-μ(t)) 2 +w n (t)×(μ n (t)-μ(t)) 2 (14)

[0136] Among them, w s (t) represents the proportion of “signal” photons; t represents the number of “signal” photons; N represents the number of photons in the water body photon data after one filtering; w n (t) represents the proportion of “noise” photons; μ s (t) represents the mean density of the “signal” photons; den(i) represents the first density value of the i-th photon in the second elevation segment Sk; μ n (t) represents the mean density of “noise” photons; μ(t) represents the mean density of water photon data after one filtering; (σ(t)) 2 Represents the between-class density variance of “signal photons” and “noise” photons.

[0137] According to the descending order of the first density values ​​of all photons in each second elevation segment, the value of t starts from 1 and increases by 1 each time until t is N. When t = 1, den(t) is the first density value of the first photon in the second elevation segment. According to formulas (9) to (14), the inter-class density variance when t = 1 can be calculated; when t = 2, den(t-1) and den(t) are the first density values ​​of the first and second photons in the second elevation segment. According to formulas (9) to (14), the inter-class density variance when t = 2 can be calculated; and so on, according to formulas (9) to (14), the inter-class density variance when t is 1 to N can be calculated.

[0138] The maximum inter-class density variance is determined from the inter-class density variances when t is 1 to N, thereby determining t corresponding to the maximum inter-class density variance. At this time, the first density value den(t) of the tth photon in the second elevation segment is the initial density threshold of the second elevation segment.

[0139] Step S8.3: Calculate the second density threshold of the corresponding second elevation segment according to the initial density threshold of each second elevation segment and the initial density threshold of its adjacent second elevation segment. The specific formula is:

[0140]

[0141] Among them, D T2 Indicates the second elevation segment S k The second density threshold, Indicates the second elevation segment S k The initial density threshold, Indicates the second elevation segment S k The two adjacent second elevation segments S k-1 , S k+1 The initial density threshold.

[0142] Usually, the second density of photons and the second density threshold are calculated by establishing a linear relationship between water depth and photons. However, as the water depth increases, the size of the elliptical filter kernel also increases, which will cause the noise in the deep water in the water body photon data to reach the equivalent density of the signal photons in the shallow water, and under the same water depth, the density of signal photons and noise photons presents diversity. Therefore, the linear relationship between water depth and photons is used for processing, ignoring the variability and diversity of photon spatial distribution, resulting in missed detection and false detection problems. Since it is impossible to use a formula to express the nonlinear relationship between the unknown photon spatial distribution and water depth, the present invention segments the water depth (i.e., the second elevation segmentation), calculates the density threshold (i.e., the second density threshold) according to the signal photon density and the noise photon density in each water depth segment, establishes a nonlinear relationship between photon density and water depth, can better reflect the density difference between signal photons and noise photons in each water depth segment, improves the accuracy of filtering, and improves missed detection and false detection problems.

[0143] Step S9: Calculate the third density threshold according to the updated major semi-axis.

[0144] Considering that the elliptical filter kernel increases with the increase of water depth, noise photons in deep water may be introduced, and a third density threshold guided by the major semi-axis in the direction of the track distance is added. The specific formula is:

[0145]

[0146] Among them, D T3 represents the third density threshold, int represents the rounding function, N sm Indicates that in the direction along the track d sm The number of sea surface photons contained in a window of meters. According to experience, d sm Set to 1000, N sm Set to 500.

[0147] Step S10: performing secondary filtering on the water body photon data after the primary filtering according to the second density threshold, the third density threshold and the second density value of each photon to obtain seabed topography photon data.

[0148] If the second density value of the photon in the water body photon data after the first filtering is not less than the corresponding second density threshold and third density threshold, the photon is a signal photon; otherwise, it is a noise photon. Filter out the noise photons from the water body photon data after the first filtering to obtain the seabed topography photon data, such as Figure 8 shown.

[0149] Step S11: using a bidirectional seed point growth algorithm to supplement the seabed topography photon data to obtain the final seabed topography photon data.

[0150] Considering the geographical differences of photons in water body photon data, the density distribution of photons in the same water depth segment in the along-track direction is different, which leads to the inability to completely extract the seabed topography photons of different densities in the same water depth segment (that is, some signal photons are filtered out, resulting in discontinuous seabed topography photon data), which in turn leads to the problem of missed detection in the seabed topography photon data obtained by secondary filtering. In order to solve this problem, based on the prior knowledge of the continuity of the spatial distribution of the seabed topography, a bidirectional seed point growth algorithm is used to supplement the seabed topography photon data to obtain complete and continuous seabed topography photon data.

[0151] In a specific embodiment of the present invention, a bidirectional seed point growth algorithm is used to supplement the seabed topography photon data, including:

[0152] Step S11.1: Calculate the difference in along-track distance between every two adjacent photons in the seabed topography photon data. The specific formula is:

[0153] Δd ij =|d j -d i | (17)

[0154] Where, Δd ij represents the difference in the distance along the track between two adjacent photons i and j, d j represents the distance along the track of photon j, d i represents the distance along the track of photon i.

[0155] Step S11.2: Calculate the second distance threshold according to the difference between the along-track distances of every two adjacent photons. The specific formula is:

[0156] T d2 =mean i≠j (Δd ij )+stb i≠j (Δd ij ) (18)

[0157] Among them, T d2 Represents the second distance threshold, mean i≠j (Δd ij ) represents the mean of the difference in along-track distances between all two adjacent photons i and j in the seafloor topography photon data, stb i≠j (Δd ij ) represents one standard deviation of the difference in along-track distances between all two adjacent photons i and j in the seabed topography photon data, mean represents the mean function, and stb represents the one standard deviation function.

[0158] Step S11.3: Determine the discontinuous region according to the difference in along-track distances between every two adjacent photons and a second distance threshold.

[0159] When the difference Δd between the distances along the track of two adjacent photons i and j is ij Greater than the second distance threshold T d2 When , there is a discontinuous region between photons i and j, such as Fig. 9 As shown in the figure. For the seabed topography photon data, there are no other photons between photons i and j; for the water body photon data after one filtering, there are other photons between photons i and j. The bidirectional seed point growth algorithm is used to add the photons that meet the conditions between photons i and j to the seabed topography photon data to solve the problem of missed detection.

[0160] Step S11.4: For each discontinuous region, one of the two adjacent photons, photon i, is used as the initial seed point, and a KD tree search algorithm is used to search for the nearest neighbor photon in the direction of the other photon j to obtain a candidate photon;

[0161] Step S11.5: Determine whether the candidate photon is a signal photon or a noise photon; when the candidate photon is a signal photon, add the signal photon to the seabed topography photon data, and use the signal photon as a new seed point, repeat steps S11.4 and S11.5 until the seed point exceeds the discontinuous area; when the candidate photon is a noise photon, complete the processing of one side (left side) of the discontinuous area, such as Fig.10 As shown;

[0162] Step S11.6: For each discontinuous region, take the other photon j of the two adjacent photons as the initial seed point, and use the KD tree search algorithm to search for the nearest neighbor photon in the direction of one of the photons i to obtain the candidate photons;

[0163] Step S11.7: Determine whether the candidate photon is a signal photon or a noise photon; when the candidate photon is a signal photon, add the signal photon to the seabed topography photon data, and use the signal photon as a new seed point, repeat steps S11.6 and S11.7 until the seed point exceeds the discontinuous area; when the candidate photon is a noise photon, complete the processing of the other side (right side) of the discontinuous area, such as Fig.11 shown.

[0164] In step S11.5 or step S11.7, when σ2≤1, the alternative photon is a signal photon; when σ2>1 and the elevation difference between the alternative photon and the seed point is less than or equal to the first distance threshold, the alternative photon is a signal photon; when σ2>1 and the elevation difference between the alternative photon and the seed point is greater than the first distance threshold, the alternative photon is a noise photon.

[0165] Step S11.8: The seafloor topography photon data obtained in step S11.5 and step S11.7 are combined, and the combined seafloor topography photon data (such as Fig.12The track is divided into multiple sections to obtain multiple second track sections;

[0166] Step S11.9: For each second track segment, divide it into a plurality of third elevation segments according to the elevation (such as Fig.13 As shown), count the number of photons in each third elevation segment, and find the third elevation segment with the largest number of photons;

[0167] Step S11.10: If the number of photons in the third elevation segment with the largest number of photons is greater than twice the photon data of its two adjacent third elevation segments, the third elevation segment with the largest number of photons is retained, and the two adjacent third elevation segments of the third elevation segment with the largest number of photons are discarded;

[0168] If the number of photons in the third elevation segment with the largest number of photons is less than or equal to twice the photon data of its two adjacent third elevation segments, the width of the third elevation segment is reduced and the process proceeds to step S11.9 until a third elevation segment is found whose number of photons is greater than twice the photon data of its two adjacent third elevation segments.

[0169] When the width of the third elevation segment is less than 0.2 m, and a third elevation segment having a photon number greater than twice the photon data of its two adjacent third elevation segments is not found, the second along-track segment is discarded.

[0170] Step S11.11: Merge all the second along-track segments obtained in step S11.10 to obtain the final seafloor topography photon data, such as Fig.14 shown.

[0171] Considering that each water depth segment has signal photons and noise photons of comparable density, the density threshold of photons in the water depth segment filters out some signal photons in order to accurately filter out noise, making the extracted seabed topography photons discontinuous. A bidirectional seed point growth algorithm is used to supplement the discontinuous areas in the seabed topography photon data obtained by two filterings to obtain more complete, continuous and accurate seabed topography photons.

[0172] The method of the present invention is used to extract the seabed topography photon data from 4 ICESat-2 / ATL03 photon data, such as Figures 15 to 18 As shown. Figures 15 to 18 It can be seen that for the four ICESat-2 / ATL03 photon data, the present invention can obtain accurate, complete and continuous seabed topography signal photon results.

[0173] Table 1 Comparison of the present invention with three existing filtering methods

[0174]

[0175] In Table 1, the R index represents the recall rate, the P index represents the precision rate, and the F index represents the balance index between the recall rate and the precision rate. The larger the balance index, the better the extraction effect.

[0176] The method of the present invention is compared with three filtering methods, namely AVEBM (Adaptive Variable Ellipse Filtering Bathymetric Method), bimodal Gaussian fitting and quadtree segmentation, as shown in Table 1. As shown in Table 1, the F index value of the extraction result of the method of the present invention is the highest, indicating that the extraction effect of the seabed topography photon data obtained by the method of the present invention is the best.

[0177] The present invention adopts the same adaptive filter as the AVEBM method, but AVEBM introduces some noise photons around the seabed topography, resulting in insufficient filtering; the bimodal Gaussian fitting method encounters a sparse distribution of topographic signal photons due to the weakening effect of water on the laser in deep water areas, which will introduce more noise photons, and also performs poorly for situations with large terrain fluctuations; the result of the quadtree segmentation method is obviously manifested as a filtering wave, which only retains high-density signal photons, filters out low-density signal photons, and also retains high-density noise photons, resulting in sparse and discrete filtering results. The present invention significantly avoids the above-mentioned shortcomings, performs elevation segmentation on the water depth, and the adaptive density threshold (i.e., the second density threshold) is a fitting process based on the nonlinear relationship between water depth and photon density, which improves the accuracy of identifying and extracting seabed topography photons with diverse density distributions under different water depth segments, and uses a seed point growth algorithm to ensure the integrity of the seabed topography, and finally extracts accurate, complete, and continuous seabed topography photons.

[0178] Example 2

[0179] An embodiment of the present invention also provides an electronic device, which includes: a memory, a processor, and a computer program / instructions stored in the memory, and the processor executes the computer program / instructions to implement the seabed topography photon extraction method in the embodiment of the present application.

[0180] Although not shown, the electronic device includes a processor, which can perform various appropriate operations and processes according to the program and / or data stored in the read-only memory (ROM) or the program and / or data loaded from the storage portion into the random access memory (RAM). The processor can be a multi-core processor, or it can include multiple processors. In some embodiments, the processor can include a general main processor and one or more special coprocessors, such as a central processing unit, a graphics processing unit (GPU), a neural network processor (NPU), a digital signal processor (DSP), etc. In RAM, various programs and data required for device operation are also stored. The processor, ROM and RAM are connected to each other via a bus. The input / output (I / O) interface is also connected to the bus.

[0181] The processor and memory are used together to execute the program / instructions stored in the memory. When the program / instructions are executed by the computer, the methods, steps or functions described in the above embodiments can be implemented.

[0182] Although not shown, an embodiment of the present invention further provides a computer-readable storage medium having a computer program / instruction stored thereon, which, when executed by a processor, implements the method for extracting photons from the seabed topography in the embodiment of the present application.

[0183] Storage media in embodiments of the present invention include permanent and non-permanent, removable and non-removable items that can be implemented by any method or technology to store information. Examples of storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission medium that can be used to store information that can be accessed by a computing device.

[0184] Readable storage media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include temporary computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0185] Although not shown, an embodiment of the present invention further provides a computer program product, including: a computer program / instruction, which, when executed by a processor, implements the method for extracting photons from the seabed topography in the embodiment of the present application.

[0186] What is disclosed above is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or modifications within the technical scope disclosed in the present invention, which should be covered within the protection scope of the present invention.

Claims

1. A method for extracting photons from seabed topography, characterized in that: The extraction method comprises: Preprocess ICESat-2 / ATL03 photon data to obtain shallow sea photon data; Extracting sea surface photon data and water body photon data from the shallow sea section photon data; A mathematical model of an elliptical filter kernel is constructed according to the along-track distance and elevation of the photons, and the major semi-axis and the minor semi-axis of the elliptical filter kernel are initialized; Calculate a first density value of each photon in the water body photon data according to the initial values ​​of the major semi-axis and the minor semi-axis and a mathematical model of an elliptical filter kernel; Determining a first density threshold according to the water body photon data and initial values ​​of the major semi-axis and the minor semi-axis; Filtering the water body photon data once according to the first density threshold and the first density value of each photon; The major semi-axis and the minor semi-axis are updated according to the elevation of each photon in the water body photon data after the primary filtering; the second density value of the corresponding photon in the water body photon data after the primary filtering is calculated according to the updated major semi-axis and minor semi-axis and the mathematical model of the elliptical filter kernel; Calculating a second density threshold for each second elevation segment based on the first density value of each photon; Calculate the third density threshold based on the updated major semi-axis; The water body photon data after the primary filtering is subjected to secondary filtering according to the second density threshold, the third density threshold and the second density value of each photon to obtain seabed topography photon data.

2. The method for extracting photons from seabed topography according to claim 1, characterized in that: Extracting sea surface photon data and water body photon data from the shallow sea section photon data includes: The width of the first elevation segment is set, the photon data of the shallow sea segment is projected according to the elevation, and the number of photons in each first elevation segment is counted to form an elevation distribution histogram; The elevation distribution histogram is fitted using a bimodal Gaussian distribution fitting algorithm, and then the sea surface photon data and the water body photon data are extracted.

3. The method for extracting photons from seabed topography according to claim 1, characterized in that: Initializing the major semi-axis and the minor semi-axis of the elliptical filter kernel according to the sea surface photon data includes: Dividing the sea surface photon data into n segments along the track according to the step length, and calculating the maximum elevation difference of each first segment along the track; Calculate the initial values ​​of the major and minor axes of the elliptical filter kernel. The specific formula is: b0=|H max -H min |; a0=R ab b0, Among them, b0 represents the initial value of the minor semi-axis of the elliptical filter kernel, H max represents the upper boundary of the sea surface photon data, H min represents the lower boundary of the sea surface photon data, a0 represents the initial value of the major semi-axis of the elliptical filter kernel, R ab represents the shape adjustment factor of the elliptical filter kernel, Δd represents the step size, ΔSH i Indicates the maximum elevation difference of the i-th first segment along the track.

4. The method for extracting photons from seabed topography according to claim 1, characterized in that: The specific process of calculating the first density value or the second density value of each photon includes: Calculate the distance between each photon and other photons according to the mathematical model of the elliptical filter kernel; A first distance threshold is set; for each photon, the number of photons whose distance does not exceed the first distance threshold is counted, and the number of photons is the first density value or the second density value of the photon.

5. The method for extracting photons from seabed topography according to claim 1, characterized in that: The formula for determining the first density threshold is: Among them, D T1 represents the first density threshold, SN1 represents the signal photon distribution density within the initial unit elliptical kernel area, SN2 represents the noise photon distribution density within the initial unit elliptical kernel area, a0 represents the initial value of the major semi-axis of the elliptical filter kernel, b0 represents the initial value of the minor semi-axis of the elliptical filter kernel, M represents the number of photons in the water body photon data, ΔH represents the vertical distance between the highest photon and the lowest photon in the water body photon data, ΔL represents the difference between the maximum along-track distance and the minimum along-track distance in the water body photon data, m l Represents the number of photons within the elevation range of the noise photon distribution in the water body photon data, N H Indicates the elevation range of the noise photon distribution in the set water body photon data.

6. The method for extracting photons from seabed topography according to claim 1, characterized in that: The update formulas of the major and minor semi-axis are: b=b0+τ|h p -H min |,a=R ab b; Among them, b represents the updated minor semi-axis, b0 represents the initial value of the minor semi-axis of the ellipse filter kernel, τ represents the scaling factor of the ellipse, and h p represents the height of the pth photon, H min represents the lower boundary of the sea surface photon data, a represents the updated major semi-axis, and R ab Represents the shape adjustment factor of the elliptical filter kernel; Preferably, calculating the second density threshold in each second elevation segment according to the first density value of each photon comprises: The width of the second elevation segment is set, the water body photon data after the primary filtering is projected according to the elevation, the first density values ​​of all photons in each second elevation segment are counted, and the first density values ​​are arranged in descending order; The maximum inter-class variance method is used to calculate the initial density threshold of each second elevation segment; The second density threshold of the corresponding second elevation segment is calculated according to the initial density threshold of each second elevation segment and the initial density thresholds of its adjacent second elevation segments.

7. The method for extracting photons from seabed topography according to any one of claims 1 to 6, characterized in that: The extraction method also includes using a bidirectional seed point growth algorithm to supplement the seabed topography photon data, specifically including: Calculating the difference in along-track distances between every two adjacent photons in the seabed topography photon data; Calculate the second distance threshold according to the difference between the along-track distances of every two adjacent photons; Determine the discontinuous area according to the difference in along-track distances between every two adjacent photons and a second distance threshold; For discontinuous areas, one of the two adjacent photons is used as the initial seed point, and the nearest neighbor photon search is performed in the direction of the other photon to obtain the candidate photon; Determine whether the candidate photon is a signal photon or a noise photon; when the candidate photon is a signal photon, add the signal photon to the seabed topography photon data, and use the signal photon as a new seed point, repeat the steps of nearest neighbor photon search and candidate photon judgment until the seed point exceeds the discontinuous area; when the candidate photon is a noise photon, complete the processing of one side of the discontinuous area; Taking the other photon of the two adjacent photons as the initial seed point, a nearest neighbor photon search is performed in the direction of one of the photons to obtain a candidate photon; Determine whether the candidate photon is a signal photon or a noise photon; when the candidate photon is a signal photon, add the signal photon to the seabed topography photon data, and use the signal photon as a new seed point to repeat the steps of nearest neighbor photon search and candidate photon judgment until the candidate photon exceeds the discontinuous area; when the candidate photon is a noise photon, complete the processing of the other side of the discontinuous area; Merging the seabed topography photon data obtained by processing both sides of the discontinuous area, and dividing the merged seabed topography photon data into multiple segments along the track to obtain multiple second along-track segments; For each of the second along-track segments, multiple third elevation segments are obtained according to the elevation segmentation, the number of photons in each third elevation segment is counted, and the third elevation segment with the largest number of photons is found; Determine whether the number of photons in the third elevation segment with the largest number of photons is greater than twice the photon data of its two adjacent third elevation segments. If so, retain the third elevation segment with the largest number of photons and discard the two adjacent third elevation segments of the third elevation segment with the largest number of photons; if not, reduce the width of the third elevation segment, and repeat the steps of segmenting by elevation and determining until a third elevation segment with a number of photons greater than twice the photon data of its two adjacent third elevation segments is found; All the processed second along-track segments are merged to obtain the final seabed topography photons.

8. An electronic device comprising a memory, a processor, and a computer program / instruction stored in the memory, characterized in that: The processor executes the computer program / instructions to implement the method for extracting photons from seabed topography according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instructions are executed by a processor, the method for extracting photons from seabed topography as claimed in any one of claims 1 to 7 is implemented.

10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instruction is executed by a processor, the method for extracting photons from seabed topography as claimed in any one of claims 1 to 7 is implemented.

Images