Method and device for extracting ground photons under mountain forest and electronic equipment

Abstract

The invention relates to the laser radar processing field, and provides a mountain forest ground photon extraction method comprising the following steps: obtaining original photon counting laser radar data, and carrying out denoising processing on the original photon data to obtain signal photon point cloud data; performing transverse partitioning on the signal photon point cloud data, rotating the signal photon to an approximate level, establishing an elevation frequency histogram, and performing fitting to obtain pseudo waveform data; ground wave crests are extracted from the pseudo waveform data, photons in the ground wave crests are extracted and reversely rotated, and first ground photons are obtained; constructing a search neighborhood for the first ground photons to obtain a photon characteristic value, a photon maximum density and a photon maximum density direction, and rejecting near-ground non-ground photons by using the characteristic value to obtain second ground photons; constructing spatial distribution characteristic parameters, and extracting ground photon seed points in a data window based on the spatial distribution characteristic parameters, the photon maximum density and the photon maximum density direction; and encrypting the ground photon seed points to obtain complete under-forest ground photons.

Description

Technical Field

[0001] This invention relates to the field of lidar processing, and more particularly to a ground-based photon extraction method, apparatus, and electronic device. Background Technology

[0002] Topographic data, as a crucial foundation for the geographic information mapping industry, is a key factor in resource management, disaster risk assessment, and policy formulation. Spaceborne lidar measures the spatial location and topographic features of surface targets by transmitting and receiving laser pulse signals. The next-generation ice, cloud, and ground elevation satellite (ICESat-2) employs a 532nm wavelength photon counting radar, boasting advantages such as high repetition rate (10kHz) and multiple beams (three groups of six beams). It can acquire small-spot, high-density photon point cloud data, providing a data foundation for all-weather, large-scale, high-precision topographic surveys.

[0003] Since mountainous areas are often accompanied by complex terrain conditions and are affected by the impact of vegetation cover on laser penetration, the extraction of ground photons under mountain forests has the following two major challenges: (1) In areas with complex terrain or dense vegetation cover, there may be a sparse distribution of ground photons, making it difficult to obtain ground photons completely and accurately; (2) In areas with complex vertical vegetation structure, noise photons remaining near the ground surface and low canopy photons near the ground surface can easily interfere with the extraction of ground photons.

[0004] Currently, no effective solution has been proposed for the precise acquisition of ground photons. Summary of the Invention

[0005] This application provides a method, apparatus, and electronic device for extracting ground photons under mountain forests. Based on pseudo-waveform and spatial distribution characteristic parameters, it achieves accurate identification of ground photons under forests, thereby effectively extracting forest terrain information in mountainous areas and overcoming the problems of difficult extraction of sparse ground photons and interference from non-ground photons under complex terrain.

[0006] Firstly, this application provides a method for extracting ground photons under forest cover in mountainous areas. The method includes: acquiring raw photon-counting lidar data; denoising the raw photon data to obtain signal photon point cloud data; wherein the raw photon-counting lidar data includes at least one segment of forest / mountain data; dividing the signal photon point cloud data into blocks along the orbital direction, rotating the signal photons to an approximate horizontal position based on a fitted approximate slope, establishing an elevation-frequency histogram, and using spline fitting to obtain pseudo-waveform data; extracting ground photon peaks from the pseudo-waveform data, and extracting signal photon point clouds within the ground photon peaks based on approximate slope information. The first ground photon is obtained by inverse rotation. An adaptive elliptical search neighborhood is constructed for the first ground photon to obtain at least one set of first ground photon feature values, maximum photon density, and maximum photon density direction. Near-ground non-ground photons are removed using the feature values ​​to obtain the second ground photon. Spatial distribution feature parameters are constructed. Based on the second ground photon, the spatial distribution feature parameters, maximum photon density, and maximum photon density direction are used to extract ground photon seed points in the data window. The ground photon seed points are used as the initial ground photons, and the second ground photon is used as the photon to be encrypted. The adaptive TIN model is used to encrypt the ground photons under the mountain forest to obtain the ground photons.

[0007] In some possible implementations, the raw photon-counting lidar data is denoised to obtain signal photon point cloud data, including: dividing the raw photon-counting lidar data into horizontal blocks and calculating the average elevation value of the photon point cloud data within each block. Standard deviation of elevation Remove elevation in Remove noisy photons; calculate the core distance and reachable distance of the photon point cloud data, and calculate the optimal reachable distance threshold based on the maximum inter-class difference method to remove noisy photons; recalculate the mean elevation of the photon point cloud data. Standard deviation of elevation Remove elevation in The residual noise photons outside are used to obtain signal photon point cloud data.

[0008] In some possible implementations, the signal photon point cloud data is divided into blocks along the track direction, and the signal photons are rotated to an approximate horizontal position based on the fitted approximate slope. An elevation-frequency histogram is then established, and pseudo-waveform data is obtained using spline fitting. This includes: estimating the approximate slope of each data block using a linear regression method; rotating the signal photon point cloud data to an approximate horizontal position according to the approximate slope center to obtain rotated signal photon point cloud data; statistically analyzing the elevation values ​​and establishing an elevation-frequency histogram based on the rotated signal photon point cloud data within each block; and interpolating the elevation-frequency histogram based on cubic spline fitting to obtain pseudo-waveform data of the signal photon point cloud data within each data block.

[0009] In some possible implementations, ground photon peak extraction is performed on the pseudo-waveform data, and the signal photon point cloud within the ground peak is inversely rotated according to the approximate slope information to obtain the first ground photon. This includes: decomposing the pseudo-waveform data into high-frequency detail signals and low-frequency approximate signals using wavelet transform; identifying the specific elevation value of the peak where the ground photon is located based on the low-frequency approximate signal; constructing an elevation buffer on both sides of the elevation value to extract the signal photon; and inversely rotating according to the approximate slope information to obtain the first ground photon.

[0010] In some possible implementations, an adaptive elliptical search neighborhood is constructed for the first ground photons, and at least one set of eigenvalues, maximum photon density, and maximum photon density direction for the first ground photons are obtained. Near-ground non-ground photons are then eliminated using the eigenvalues ​​to obtain the second ground photons. This includes: constructing an adaptive elliptical search neighborhood centered on at least one set of first ground photons; rotating the search neighborhood to include the maximum number of first ground photons within it; obtaining the maximum photon density value and maximum density direction; and constructing a covariance matrix using the coordinate information of all first ground photons within the search neighborhood and calculating eigenvalues. and ,and < The eigenvalues ​​of the central photon in the elliptical search neighborhood are obtained; the eigenvalues ​​of the first ground photon are then used. and The ratio is compared with the filtering threshold, and near-ground non-ground photons are removed based on the comparison result to obtain the second ground photon.

[0011] In some possible implementations, spatial distribution feature parameters are constructed, and ground photon seed points are extracted in a data window based on the second ground photon using the spatial distribution feature parameters, photon maximum density, and photon maximum density direction. This includes: constructing spatial distribution feature parameter eigenvalues ​​(SU), total variance (OM), feature entropy (EI), and linearity (LI) based on the feature values ​​of the first ground photon; dividing the data window along the orbital direction for the second ground photon, sorting the spatial distribution feature parameters of each photon within the window, and selecting photons with SU values ​​in a higher range and OM, EI, and LI values ​​in a lower range to form a non-repeating union; based on the non-repeating union, selecting the photon with the highest density in the data window whose photon maximum density direction meets a set threshold as the ground photon seed point.

[0012] In some possible implementations, ground photon seed points are used as initial ground photons, and second ground photons are used as photons to be encrypted. An adaptive TIN model is used to encrypt the complete forest understory ground photons. This includes: using ground photon seed points as initial ground photons, using second ground photons as photons to be encrypted, and using an adaptive TIN model to gradually expand the set of ground photons from the photons to be encrypted according to the distance and angle between photons, to obtain mountain forest understory ground photons.

[0013] Secondly, this application provides a ground photon extraction device for mountain forest understory, the device comprising: a processing module for acquiring raw photon counting lidar data, performing noise reduction processing on the raw photon data to obtain signal photon point cloud data; wherein the raw photon counting lidar data includes at least one segment of forest mountain data; a fitting module for dividing the signal photon point cloud data into blocks along the track direction, rotating the signal photons to approximately horizontal according to the fitted approximate slope, establishing an elevation-frequency histogram, and using spline fitting to obtain pseudo-waveform data; and a photon coarse extraction module for extracting ground photon peaks from the pseudo-waveform data, and extracting signal photons within the ground photon peaks according to the approximate slope information. The sub-point cloud is inversely rotated to obtain the first ground photon; the photon filtering module is used to construct an adaptive elliptical search neighborhood for the first ground photon, obtain at least one set of first ground photon feature values, photon maximum density, and photon maximum density direction, and use the feature values ​​to remove near-ground non-ground photons to obtain the second ground photon; the seed module is used to construct spatial distribution feature parameters, and extract ground photon seed points in the data window based on the second ground photon using the spatial distribution feature parameters, photon maximum density, and photon maximum density direction; the generation module uses the ground photon seed points as the initial ground photon and the second ground photon as the photon to be encrypted, and uses an adaptive TIN model to encrypt to obtain the mountain forest underground ground photon.

[0014] In some possible implementations, the processing module divides the raw photon-counting lidar data into horizontal blocks and calculates the average elevation of the photon point cloud data within each block. Standard deviation of elevation Remove elevation in Remove noisy photons; calculate the core distance and reachable distance of the photon point cloud data, and calculate the optimal reachable distance threshold based on the maximum inter-class difference method to remove noisy photons; recalculate the mean elevation of the photon point cloud data. Standard deviation of elevation Remove elevation in The residual noise photons outside are used to obtain signal photon point cloud data.

[0015] In some possible implementations, the fitting module uses a linear regression method to estimate the approximate slope of each data block, rotates the signal photon point cloud data to an approximate horizontal position according to the approximate slope center, and obtains the rotated signal photon point cloud data; based on the rotated signal photon point cloud data within the block, the elevation values ​​are statistically analyzed and an elevation frequency histogram is established, and the pseudo waveform data of the signal photon point cloud data within each data block is obtained by interpolating the elevation frequency histogram based on cubic spline fitting.

[0016] In some possible implementations, the photon coarse extraction module uses wavelet transform to decompose the pseudo-waveform data into high-frequency detail signals and low-frequency approximate signals. Based on the low-frequency approximate signals, it identifies the specific elevation value of the peak where the ground photon is located, constructs an elevation buffer on both sides of the elevation value to extract the signal photon, and performs inverse rotation based on the approximate slope information to obtain the first ground photon.

[0017] In some possible implementations, the photon filtering module 54 constructs an adaptive elliptical search neighborhood centered on at least one set of first ground photons, rotates the search neighborhood to contain the most first ground photons, and obtains the maximum photon density value and the maximum density direction; it then constructs a covariance matrix using the coordinate information of all first ground photons within the search neighborhood and calculates eigenvalues. and ,and < The eigenvalues ​​of the central photon in the elliptical search neighborhood are obtained; the eigenvalues ​​of the first ground photon are then used. and The ratio is compared with the filtering threshold, and near-ground non-ground photons are removed based on the comparison result to obtain the second ground photon.

[0018] In some possible implementations, the seed module constructs spatial distribution characteristic parameters (SU), total variance (OM), characteristic entropy (EI), and linearity (LI) based on the first ground photon characteristic value. For the second ground photon, a data window is divided along the orbital direction. Within the window, the spatial distribution characteristic parameters of each photon are sorted, and photons with SU values ​​in a higher range and OM, EI, and LI values ​​in a lower range are selected to form a non-repeating union. Based on the non-repeating union, the photon with the highest density in the data window whose maximum density direction meets a set threshold is selected as the ground photon seed point.

[0019] In some possible implementations, the generation module uses ground photon seed points as initial ground photons, takes second ground photons as photons to be encrypted, and uses an adaptive TIN model to gradually expand the set of ground photons from the photons to be encrypted according to the distance and angle between photons, thus obtaining mountain forest underground ground photons.

[0020] Thirdly, this application provides a computing device, comprising: at least one memory for storing a program; and at least one processor for executing the program stored in the memory, wherein when the program stored in the memory is executed, the processor is configured to execute a method as described in any of the first aspects.

[0021] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when run on a processor, causes the processor to perform the method as described in any of the first aspects. Attached Figure Description

[0022] To more clearly illustrate the technical solutions of the various embodiments disclosed in this specification, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only a few embodiments disclosed in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] The accompanying drawings used in the description of the embodiments or prior art are briefly introduced below.

[0024] Figure 1 This is a flowchart of a method for extracting ground photons in mountain forests, as disclosed in an embodiment of this application.

[0025] Figure 2 The image shows the results of coarse extraction of ground photons provided in the example.

[0026] Figure 3 This is a result diagram of ground photon seed points provided in an embodiment of this application;

[0027] Figure 4 This application provides a result image of the complete encrypted forest understory photons for an embodiment of the present application;

[0028] Figure 5 This application provides a schematic diagram of a ground photon extraction device for mountain forests;

[0029] Figure 6 This application provides a computing device in its embodiments. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions in the embodiments of this application will be described below with reference to the accompanying drawings.

[0031] In the description of the embodiments of this application, the words "exemplary," "for example," or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "exemplary," "for example," or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the words "exemplary," "for example," or "for instance" is intended to present the relevant concepts in a specific manner.

[0032] In the description of the embodiments in this application, the term "and / or" is merely a description of the association relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, B existing alone, and A and B existing simultaneously. Furthermore, unless otherwise stated, the term "multiple" means two or more. For example, multiple systems refer to two or more systems, and multiple terminals refer to two or more terminals.

[0033] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. The terms "comprising," "including," "having," and their variations all mean "including but not limited to," unless otherwise specifically emphasized.

[0034] In the description of the embodiments in this application, "some embodiments" are mentioned, which describe a subset of all possible embodiments. However, it is understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict.

[0035] In the description of the embodiments of this application, the terms "first, second, third, etc." or module A, module B, module C, etc. are used only to distinguish similar objects and do not represent a specific ordering of objects. It is understood that, where permitted, a specific order or sequence can be interchanged so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0036] In the description of the embodiments of this application, the reference numerals for the steps, such as S110, S120, etc., do not necessarily indicate that the steps will be executed in this manner. Where permissible, the order of the steps can be interchanged or executed simultaneously.

[0037] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0038] Figure 1 This is a flowchart illustrating a method for extracting ground photons in mountainous forest understory, as disclosed in an embodiment of this application. Figure 1 As shown, it includes the following steps:

[0039] In step S101, the raw photon counting lidar data is acquired, and the raw photon data is denoised to obtain signal photon point cloud data. The raw photon counting lidar data includes at least one segment of forest / mountain terrain data.

[0040] Specifically, the original photon counting lidar data is first divided into horizontal blocks, and the average elevation value of the photon point cloud data within each block is calculated. Standard deviation of elevation Remove elevation in Obvious noise photons outside.

[0041] Using the improved OPTICS density clustering method, the core distance and reachability of photon point cloud data are calculated, and the optimal reachability threshold is calculated based on the maximum inter-class difference method to remove noisy photons.

[0042] Recalculate the mean elevation of the photon point cloud data Standard deviation of elevation Remove elevation in The residual noise photons outside are used to obtain signal photon point cloud data.

[0043] In some possible embodiments, the removal of residual noise may be omitted as appropriate.

[0044] In step S102, the signal photon point cloud data is divided into horizontal blocks along the track direction, and the signal photons are rotated to an approximate horizontal position according to the fitted approximate slope to establish an elevation frequency histogram. The pseudo waveform data is obtained by fitting with cubic splines.

[0045] Specifically, the signal photons are divided into horizontal blocks along the orbital direction, and a linear fitting method is used to obtain the approximate terrain slope of the signal photons within the data blocks and rotate the center to approximately horizontal.

[0046] In some possible embodiments, the signal photons within the block are rotated to approximately horizontal angles using formula (1) to eliminate the influence of terrain:

[0047] (1)

[0048] in, The approximate slope of signal photons within the data block. and These represent the distance and elevation of the signal photon along its trajectory after rotation. and These represent the orbital distance and elevation values ​​of the original signal photon, respectively.

[0049] Based on the rotated signal photon point cloud data within the block, the elevation values ​​are statistically analyzed and an elevation frequency histogram is established. Based on cubic spline fitting, the elevation frequency histogram is interpolated to obtain the pseudo waveform data of the signal photon point cloud data within each data block.

[0050] In step S103, wavelet transform and peak detection methods are used to extract ground photon peaks from pseudo-waveform data, and the signal photon point cloud within the ground peak is rotated inversely according to approximate slope information to obtain the first ground photon.

[0051] Specifically, the pseudo-waveform data is decomposed into high-frequency detail signals and low-frequency approximation signals using wavelet transform. Based on the low-frequency approximation signal, the first maximum peak is identified as the ground photon peak. The specific elevation value of the ground photon peak is obtained. An elevation buffer is constructed on both sides of the elevation value of the ground photon peak to extract signal photons within the elevation range. The photons are then rotated inversely according to the approximate slope information to obtain the first ground photon.

[0052] In some possible embodiments, the peaks exceeding the elevation frequency threshold of formula (2) are identified as the first maximum peaks, i.e., the ground photon peaks:

[0053] (2)

[0054] in For elevation frequency threshold, The number of peaks in the low-frequency approximation signal. For the first Peak elevation frequency, For the first The frequency of each trough elevation.

[0055] In step S104, an adaptive elliptical search neighborhood is constructed for the first ground photon, and the feature values, maximum photon density, and maximum photon density direction of each first ground photon are obtained. The feature values ​​are used to eliminate near-ground non-ground photons to obtain the second ground photon.

[0056] Specifically, an adaptive elliptical search neighborhood is constructed with each first ground photon as the center. The search neighborhood is rotated counterclockwise with a step size of 10° until it contains the most first ground photons, and the maximum photon density value is obtained. The angle between the major axis of the elliptical search neighborhood and the horizontal line is taken as the direction of the maximum photon density.

[0057] The covariance matrix is ​​constructed by using the coordinate information of all coarsely extracted signal photons in the elliptical search neighborhood, along with their distance and elevation, and then the eigenvalues ​​are calculated. , , < This serves as the eigenvalue result of the central photon.

[0058] In some possible implementations, if the number of photons in the elliptical search neighborhood is insufficient to construct the covariance matrix, the major and minor axes of the neighborhood are gradually increased in steps of 0.1 m until there are enough photons in the neighborhood to participate in the calculation.

[0059] Using the first ground photon characteristic value , The ratio is compared with the non-ground photon filtering threshold in equation (3), and near-ground non-ground photons with values ​​less than the threshold are removed to obtain the second ground photon:

[0060] (3)

[0061] in This is the non-ground photon filtering threshold. , These are characteristic values ​​in the elevation direction and along the track direction, respectively, in some possible embodiments. The value is 30.

[0062] In step S105, spatial distribution feature parameters are constructed, and ground photon seed points are extracted in the data window based on the second ground photon using the spatial distribution feature parameters, the maximum photon density, and the direction of the maximum photon density.

[0063] Specifically, based on the first ground photon eigenvalue , Based on equations (4) to (7), construct the spatial distribution characteristic parameters SU (sum of eigenvalues), OM (total variance), EI (eigenentiation), and LI (linearity):

[0064] (4)

[0065] (5)

[0066] (6)

[0067] (7)

[0068] The data window for the second ground photon along the orbital direction is divided, and the spatial distribution characteristic parameters of each photon are sorted within the window. Photons with SU values ​​in the higher range and OM, EI and LI values ​​in the lower range are selected to form a non-repeating union.

[0069] In some possible embodiments, the SU value ranges from [85%, 100%], and the OM, EI and LI values ​​range from [0, 15%].

[0070] Based on the photon set, the photon with the highest density in the data window whose direction of maximum photon density meets the set threshold is selected as the ground photon seed point.

[0071] In some possible embodiments, the width of the data window is 15 m, and the maximum density direction threshold is set as follows:

[0072] If the direction of maximum density exceeds 80°, i.e. the direction of vegetation growth, the photon is identified as a canopy photon rather than a ground photon.

[0073] If the absolute value of the difference in the direction of maximum density between adjacent ground photon seed points is greater than 20°, then the photon is considered not to satisfy the continuity of ground photons between windows.

[0074] In some possible embodiments, if the direction of maximum photon density exceeds a threshold, the search continues downwards based on the photon density to find photons that meet the conditions as ground photon seed points.

[0075] In step S106, the ground photon seed point is used as the initial ground photon, and the second ground photon is used as the photon to be encrypted. The adaptive TIN model is used to encrypt the complete forest understory ground photon.

[0076] Specifically, based on the ground photon seed point as the initial ground photon, the second ground photon is used as the photon to be encrypted. Using the adaptive TIN model, the set of encrypted ground photons is gradually expanded from the photon to be encrypted according to the distance and angle between the photons, so as to obtain the complete forest underground ground photon.

[0077] In some possible embodiments, a distance threshold needs to be set. The value is 1 m.

[0078] Figure 2 This is a result image of the first ground photon point cloud data provided in an embodiment of this application. (See image below.) Figure 2 As shown, the dark gray dots represent the first ground photons, and the light gray dots represent the signal photons.

[0079] from Figure 2 As can be seen, after the ground photon coarse extraction in steps S101-S103, canopy photons in the signal photons can be basically removed. At the same time, ground photons can still be stably extracted in areas with complex terrain or dense vegetation.

[0080] Figure 3 The image shows the result of the ground photon seed points obtained after steps S104-S105, where dark gray dots represent ground photon seed points and light gray dots represent the first ground photon.

[0081] Depend on Figure 3 As can be seen, after the ground photon seed point extraction in steps S104-S105, the ground photon seed points extracted based on the photon spatial distribution characteristic parameters are continuously distributed in the first ground photon, which overcomes the interference of near-ground canopy photons or noise photons as much as possible.

[0082] Figure 4The image shows the encrypted complete forest underground photons obtained after step S106, where dark gray dots represent encrypted forest underground photons and light gray dots represent the first ground photons.

[0083] Depend on Figure 4 As can be seen, after the ground photons are expanded and encrypted in step S106, the continuous and complete ground photon point cloud data is the mountain forest underground ground photons obtained by the method of the present invention.

[0084] The above is an introduction to the voice quality assessment method provided by the embodiments of this application. It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. In addition, in some possible implementations, each step in the above embodiments may be selectively executed according to the actual situation, and may be partially or fully executed, without limitation here. Furthermore, all or part of any feature of any of the above embodiments may be freely and arbitrarily combined without contradiction; the combined technical solution is also within the scope of this application.

[0085] Next, based on the above, the ground photon extraction device for mountain forest understory provided in the embodiments of this application will be described. For relevant descriptions of concepts, formulas, etc., involved in the following content, please refer to the above text.

[0086] Figure 5 This application provides a schematic diagram of a ground photon extraction device for mountain forest understory, as illustrated in this embodiment. Figure 5 As shown, the ground photon extraction device 50 in the mountain forest includes a processing module 51, a fitting module 52, a photon coarse extraction module 53, a photon filtering module 54, a seed module 55, and a generation module 56.

[0087] In the device, the processing module 51 acquires the raw photon counting lidar data, performs noise reduction processing on the raw photon data, and obtains signal photon point cloud data; wherein, the raw photon counting lidar data includes at least one segment of forest and mountain data.

[0088] The fitting module 52 divides the signal photon point cloud data into blocks along the track direction, rotates the signal photons to an approximate horizontal position according to the fitted approximate slope, establishes an elevation-frequency histogram, and uses spline fitting to obtain pseudo-waveform data.

[0089] The photon coarse extraction module 53 extracts the ground photon peaks from the pseudo-waveform data and rotates the signal photon point cloud within the ground photon peaks in reverse according to the approximate slope information to obtain the first ground photon.

[0090] The photon filtering module 54 constructs an adaptive elliptical search neighborhood for the first ground photon, obtains at least one set of first ground photon feature values, photon maximum density and photon maximum density direction, and uses the feature values ​​to remove near-ground non-ground photons to obtain the second ground photon.

[0091] Seed module 55 constructs spatial distribution feature parameters, and extracts ground photon seed points in the data window based on the spatial distribution feature parameters, photon maximum density, and photon maximum density direction of the second ground photon.

[0092] The generation module 56 uses the ground photon seed point as the initial ground photon and the second ground photon as the photon to be encrypted, and uses the adaptive TIN model to encrypt the ground photon under the mountain forest.

[0093] In some possible implementations, the processing module 51 divides the raw photon counting lidar data into horizontal blocks and calculates the average elevation of the photon point cloud data within each block. Standard deviation of elevation Remove elevation in Remove noisy photons; calculate the core distance and reachable distance of the photon point cloud data, and calculate the optimal reachable distance threshold based on the maximum inter-class difference method to remove noisy photons; recalculate the mean elevation of the photon point cloud data. Standard deviation of elevation Remove elevation in The residual noise photons outside are used to obtain signal photon point cloud data.

[0094] In some possible implementations, the fitting module 52 uses a linear regression method to estimate the approximate slope of each data block, rotates the signal photon point cloud data to an approximate horizontal position according to the approximate slope center, and obtains the rotated signal photon point cloud data; based on the rotated signal photon point cloud data within the block, the elevation values ​​are statistically analyzed and an elevation frequency histogram is established, and the elevation frequency histogram is interpolated based on cubic spline fitting to obtain the pseudo waveform data of the signal photon point cloud data within each data block.

[0095] In some possible implementations, the photon coarse extraction module 53 uses wavelet transform to decompose the pseudo-waveform data into high-frequency detail signals and low-frequency approximate signals. Based on the low-frequency approximate signals, it identifies the specific elevation value of the peak where the ground photon is located, constructs an elevation buffer on both sides of the elevation value to extract the signal photon, and performs inverse rotation according to the approximate slope information to obtain the first ground photon.

[0096] In some possible implementations, the photon filtering module 54 constructs an adaptive elliptical search neighborhood centered on at least one set of first ground photons, rotates the search neighborhood to contain the most first ground photons, and obtains the maximum photon density value and the maximum density direction; it then constructs a covariance matrix using the coordinate information of all first ground photons within the search neighborhood and calculates eigenvalues. and ,and < The eigenvalues ​​of the central photon in the elliptical search neighborhood are obtained; the eigenvalues ​​of the first ground photon are then used. and The ratio is compared with the filtering threshold, and near-ground non-ground photons are removed based on the comparison result to obtain the second ground photon.

[0097] In some possible implementations, the seed module 55 constructs spatial distribution characteristic parameters (SU), total variance (OM), characteristic entropy (EI), and linearity (LI) based on the first ground photon characteristic values; divides the data window for the second ground photon along the orbital direction, sorts the spatial distribution characteristic parameters of each photon within the window, and selects photons with SU values ​​in a higher range and OM, EI, and LI values ​​in a lower range to form a non-repeating union; based on the non-repeating union, selects the photon with the highest density in the data window whose maximum density direction meets a set threshold as the ground photon seed point.

[0098] In some possible implementations, the generation module 56 uses ground photon seed points as initial ground photons, takes second ground photons as photons to be encrypted, and uses an adaptive TIN model to gradually expand the set of ground photons from the photons to be encrypted according to the distance and angle between photons, so as to obtain mountain forest underground ground photons.

[0099] As an example of a software functional unit, processing module 51 may include code running on a computing instance. The computing instance may include at least one of a physical host (computing device), a virtual machine, or a container. Further, the aforementioned computing instance may be one or more. For example, processing module 51 may include code running on multiple hosts / virtual machines / containers. It should be noted that the multiple hosts / virtual machines / containers used to run the code may be distributed within the same region or in different regions. Further, the multiple hosts / virtual machines / containers used to run the code may be distributed within the same availability zone (AZ) or in different AZs, each AZ including one or more geographically proximate data centers. Typically, a region may include multiple AZs.

[0100] Similarly, multiple hosts / virtual machines / containers used to run this code can be distributed within the same Virtual Private Cloud (VPC) or across multiple VPCs. Typically, a VPC is set up within a region. Communication between two VPCs within the same region, as well as between VPCs in different regions, requires a communication gateway to be set up within each VPC to enable interconnection between VPCs.

[0101] This application also provides a computing device 70. For example... Figure 5 As shown, the computing device 70 includes a bus 72, a processor 74, a memory 76, and a communication interface 78. The processor 74, the memory 76, and the communication interface 78 communicate with each other via the bus 72. The computing device 70 can be a computing device or a terminal device. It should be understood that this application does not limit the number of processors and memories in the computing device 70.

[0102] Bus 72 can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 5 A single line may be used to represent a bus, but this does not mean that there is only one bus or one type of bus. Bus 74 may include a path for transmitting information between various components of computing device 70 (e.g., memory 76, processor 74, communication interface 78).

[0103] The processor 74 may include any one or more processors such as a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor (MP), or a digital signal processor (DSP).

[0104] The memory 76 may include volatile memory, such as random access memory (RAM). The processor 104 may also include non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid state drive (SSD).

[0105] The memory 76 stores executable program code, and the processor 74 executes the executable program code to implement the aforementioned functions respectively. Figure 5 The processing module 51 shown functions to implement all or part of the steps of the method in the above embodiments. That is, the memory 76 stores instructions for executing all or part of the steps in the method of the above embodiments.

[0106] The communication interface 78 uses transceiver modules, such as, but not limited to, network interface cards and transceivers, to enable communication between the computing device 70 and other devices or communication networks.

[0107] This application provides a computer-readable storage medium storing a computer program that, when run on a processor, causes the processor to perform the method as described in any of the first aspects.

[0108] It is understood that the processor in the embodiments of this application may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. A general-purpose processor may be a microprocessor or any conventional processor.

[0109] The method steps in the embodiments of this application can be implemented in hardware or by a processor executing software instructions. The software instructions can consist of corresponding software modules, which can be stored in random access memory (RAM), flash memory, read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), registers, hard disks, portable hard disks, CD-ROMs, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can reside in an ASIC.

[0110] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in or transmitted through a computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0111] It is understood that the various numerical designations used in the embodiments of this application are merely for descriptive convenience and are not intended to limit the scope of the embodiments of this application.

Claims

1. A method for extracting photons from the ground under forest cover in mountainous areas, characterized in that, The method includes: The raw photon counting lidar data is acquired, and the raw photon data is denoised to obtain signal photon point cloud data; wherein the raw photon counting lidar data includes at least one segment of forest and mountain data; The signal photon point cloud data is divided into blocks along the track direction, and the signal photons are rotated to an approximate horizontal position according to the fitted approximate slope to establish an elevation frequency histogram. Pseudo-waveform data is obtained by using spline fitting. Ground photon peaks are extracted from the pseudo-waveform data, and the signal photon point cloud within the ground photon peaks is inversely rotated according to the approximate slope information to obtain the first ground photon. An adaptive elliptical search neighborhood is constructed for the first ground photon, and at least one set of first ground photon feature values, photon maximum density, and photon maximum density direction are obtained. Near-ground non-ground photons are eliminated using the feature values ​​to obtain the second ground photon. Construct spatial distribution feature parameters, and extract ground photon seed points in the data window based on the second ground photon using the spatial distribution feature parameters, the maximum photon density, and the direction of the maximum photon density; Using the ground photon seed point as the initial ground photon and the second ground photon as the photon to be encrypted, the adaptive TIN model is used to encrypt the ground photon under the mountain forest.

2. The method according to claim 1, characterized in that, The denoising process of the original photon counting lidar data to obtain signal photon point cloud data includes: The original photon counting lidar data is divided into horizontal blocks, and the average elevation value of the photon point cloud data within each block is calculated. Standard deviation of elevation Remove elevation in External noise photons; Calculate the core distance and reachable distance of the photon point cloud data, and calculate the optimal reachable distance threshold based on the maximum inter-class difference method to remove noisy photons; Recalculate the mean elevation of the photon point cloud data Standard deviation of elevation Remove elevation in The residual noise photons outside are used to obtain signal photon point cloud data.

3. The method according to claim 1, characterized in that, The process involves dividing the signal photon point cloud data into blocks along the orbital direction, rotating the signal photons to an approximate horizontal position based on a fitted approximate slope, establishing an elevation-frequency histogram, and using spline fitting to obtain pseudo-waveform data, including: The approximate slope of each data block is estimated using a linear regression method. The signal photon point cloud data is then rotated to an approximate horizontal position along the approximate slope center to obtain the rotated signal photon point cloud data. Based on the rotated signal photon point cloud data within the block, the elevation values ​​are statistically analyzed and an elevation frequency histogram is established. Based on cubic spline fitting, the elevation frequency histogram is interpolated to obtain the pseudo waveform data of the signal photon point cloud data within each data block.

4. The method according to claim 1, characterized in that, The step of extracting ground photon peaks from pseudo-waveform data and inversely rotating the signal photon point cloud within the ground peaks based on approximate slope information to obtain the first ground photon includes: The pseudo-waveform data is decomposed into high-frequency detail signals and low-frequency approximation signals using wavelet transform. Based on the low-frequency approximation signals, the specific elevation value of the peak where the ground photon is located is identified. An elevation buffer is constructed on both sides of the elevation value to extract the signal photon. The approximate slope information is then reverse-rotated to obtain the first ground photon.

5. The method according to claim 1, characterized in that, The process of constructing an adaptive elliptical search neighborhood for the first ground photon, obtaining at least one set of first ground photon feature values, photon maximum density, and photon maximum density direction, and using the feature values ​​to eliminate near-ground non-ground photons to obtain the second ground photon includes: An adaptive elliptical search neighborhood is constructed with the at least one group of first ground photons as the center. The search neighborhood is rotated to contain the most first ground photons, and the maximum photon density value and the maximum density direction are obtained. Construct a covariance matrix using the coordinate information of all first ground photons within the search neighborhood and calculate the eigenvalues. and ,and < The eigenvalues ​​of the central photon in the elliptical search neighborhood are obtained. Using the characteristic value of the first ground photon and The ratio is compared with the filtering threshold, and near-ground non-ground photons are removed based on the comparison result to obtain the second ground photon.

6. The method according to claim 1, characterized in that, The construction of spatial distribution feature parameters, based on the second ground photon, uses the spatial distribution feature parameters, the maximum photon density, and the direction of the maximum photon density to extract ground photon seed points in the data window, including: Based on the first ground photon eigenvalue, spatial distribution characteristic parameters eigenvalue (SU), total variance (OM), characteristic entropy (EI), and linearity (LI) are constructed. The second ground photon is divided into data windows along the orbital direction. Within the window, the spatial distribution characteristic parameters of each photon are sorted. Photons with SU values ​​in a higher range and OM, EI, and LI values ​​in a lower range are selected to form a non-repeating union. Based on the non-repeating union, the photon with the highest density in the data window whose direction of maximum photon density meets the set threshold is selected as the ground photon seed point.

7. The method according to claim 1, characterized in that, The process of using the ground photon seed point as the initial ground photon and the second ground photon as the photon to be encrypted, and using an adaptive TIN model to encrypt the complete forest understory ground photon, includes: Based on the ground photon seed point as the initial ground photon, the second ground photon is used as the photon to be encrypted. Using the adaptive TIN model, the set of ground photons is gradually expanded from the photon to be encrypted according to the distance and angle between the photons, so as to obtain the ground photon under the mountain forest.

8. A photon extraction device for the ground under forest in mountainous areas, characterized in that, The device includes: The processing module is used to acquire raw photon counting lidar data, perform noise reduction processing on the raw photon data, and obtain signal photon point cloud data; wherein, the raw photon counting lidar data includes at least one segment of forest and mountain data; The fitting module is used to divide the signal photon point cloud data into blocks along the track direction, rotate the signal photons to an approximate horizontal position according to the fitted approximate slope, establish an elevation frequency histogram, and use spline fitting to obtain pseudo waveform data. The photon coarse extraction module is used to extract the ground photon peaks from the pseudo waveform data, and to inversely rotate the signal photon point cloud within the ground photon peaks according to the approximate slope information to obtain the first ground photon. The photon filtering module is used to construct an adaptive elliptical search neighborhood for the first ground photon, obtain at least one set of first ground photon feature values, photon maximum density and photon maximum density direction, and use the feature values ​​to remove near-ground non-ground photons to obtain the second ground photon; The seed module is used to construct spatial distribution feature parameters and extract ground photon seed points in the data window based on the second ground photon using the spatial distribution feature parameters, the maximum photon density, and the direction of the maximum photon density; The generation module uses the ground photon seed point as the initial ground photon and the second ground photon as the photon to be encrypted, and uses the adaptive TIN model to encrypt the ground photon under the mountain forest.

9. A computing device, characterized in that, include: At least one memory for storing programs; At least one processor is configured to execute a program stored in the memory, wherein when the program stored in the memory is executed, the processor is configured to perform the method as described in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program that, when executed on a processor, causes the processor to perform the method as described in any one of claims 1-7.

Images