Downlink irradiance diffuse attenuation coefficient extraction method and system

Abstract

The invention provides a downlink irradiance diffuse attenuation coefficient extraction method and system, and relates to the technical field of point cloud data processing, and the method comprises the steps: obtaining photon point cloud data of a target water body; generating photon point cloud data after depth conversion according to elevation data of photons in the photon point cloud data, determining exposure time of a horizontal pixel column, and performing three-dimensional voxelization counting processing and background removal processing on the photon point cloud data after depth conversion to obtain voxel volume and net photon count of voxels in a three-dimensional voxel grid; performing normalization processing on net photon counting through a geometry / system gain model and a Snell law to obtain a normalized counting density of voxels; and determining a downlink irradiance diffuse attenuation coefficient field according to the normalized counting density. According to the method, through depth conversion, exposure time correction, three-dimensional voxelization and background removal processing, geometric / system gain modeling and normalization processing, the acquisition precision of the downlink irradiance diffuse attenuation coefficient is remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of point cloud data processing technology, and more specifically, to a method and system for extracting the diffuse attenuation coefficient of downlink irradiance. Background Technology

[0002] Downward irradiance diffuse attenuation coefficient is a key parameter for measuring the ability of water bodies to attenuate visible light, and it is related to transparency, euphotic depth, and water quality ecological assessment. Waveform airborne lidar is typically used to obtain the downward irradiance diffuse attenuation coefficient, forming a coefficient field.

[0003] In related technologies, the acquisition of downlink irradiance diffuse attenuation coefficients mostly depends on complete echo waveforms or continuous signals. However, airborne photon counting lidar outputs discrete photon point clouds rather than continuous waveform signals. Therefore, when processing discrete photon point clouds, it is difficult to effectively isolate the influence of external factors such as exposure, geometry, system, and volume, which leads to an increase in systematic deviations and uncertainties in the extracted downlink irradiance diffuse attenuation coefficients. Summary of the Invention

[0004] The problem addressed by this invention is how to improve the accuracy of obtaining the diffuse attenuation coefficient of downlink irradiance.

[0005] To address the above problems, this invention provides a method and system for extracting the diffuse attenuation coefficient of downlink irradiance.

[0006] In a first aspect, the method for extracting the diffuse attenuation coefficient of downlink irradiance according to the present invention includes: The target water body is scanned in a circular manner by a photon counting lidar to obtain the photon point cloud data of the target water body; Based on the elevation data of each photon in the photon point cloud data, the depth data corresponding to each photon is determined, and the depth-converted photon point cloud data is generated based on the depth data. Based on the geometric characteristics of the circular scan, the exposure time of the horizontal pixel column corresponding to each photon is determined; The depth-converted photon point cloud data is subjected to three-dimensional voxelization to obtain a three-dimensional voxel grid of the depth-converted photon point cloud data. Counting and background removal processing are performed on the three-dimensional voxel grid to obtain the voxel volume and net photon count of each voxel in the three-dimensional voxel grid. The geometric / system gain of each voxel is determined by a geometric / system gain model, and the geometric flattening factor is determined according to Snell's law and the preset water refractive index. The net photon count is normalized based on the exposure time, the geometry / system gain, the voxel volume, and the geometry flattening factor to obtain the normalized count density of each voxel. The diffuse attenuation coefficient field of the downstream irradiance of the target water body is determined based on the normalized count density.

[0007] Optionally, the step of determining the depth data corresponding to each photon based on the elevation data of each photon in the photon point cloud data, and generating depth-converted photon point cloud data based on the depth data, includes: The space corresponding to the target water body is divided into multiple spatial grid blocks according to a preset size; By performing equal-width histogram statistics on the elevation data of each photon within the spatial grid block, the mode bin of the elevation distribution of the photon within each spatial grid block is obtained; The water surface elevation of the spatial grid block is determined based on the midpoint of the mode bin of the spatial grid block; Based on the water surface elevation, the elevation data of each photon is converted into the depth data; The elevation data in the photon point cloud data is replaced with the depth data to generate depth-converted photon point cloud data.

[0008] Optionally, determining the exposure time of the horizontal pixel column corresponding to each photon based on the geometric characteristics of the circular scan includes: Based on the photon point cloud data after depth conversion, the horizontal region of the target water body is divided into multiple horizontal pixel pillars according to a preset pixel size, wherein each horizontal pixel pillar corresponds to multiple photons; Based on the geometric characteristics of the circular scan and the trajectory parameters of the photon counting lidar, the exposure time of each horizontal pixel column is obtained by discrete accumulation calculation.

[0009] Optionally, the step of performing three-dimensional voxelization on the depth-converted photon point cloud data to obtain a three-dimensional voxel grid of the depth-converted photon point cloud data, and performing counting and background removal processing on the three-dimensional voxel grid to obtain the voxel volume and net photon count of each voxel in the three-dimensional voxel grid, includes: Based on the spatial distribution of the depth-converted photon point cloud data, three-dimensional voxelization is performed to construct the three-dimensional voxel mesh. Obtain the photon count within each voxel falling into the three-dimensional voxel grid and the voxel volume of each voxel; The background count rate of each horizontal pixel bar is determined based on the signal within the pulse time window; Based on the background count rate and the preset effective sampling duration of the voxel layer, the expected background count of the voxels is obtained. The net photon count of each voxel is obtained based on the difference between the photon count of each voxel and the expected background count of the voxel.

[0010] Optionally, determining the geometric / system gain of each voxel using a geometric / system gain model, and determining the geometric flattening factor based on Snell's law and a preset water refractive index, includes: Obtain the system parameters and optical characteristic parameters of the photon counting lidar; Based on the system parameters and the optical characteristic parameters, a geometric / system gain model is constructed, and the geometric / system gain of each voxel is determined through the geometric / system gain model. Based on the preset refractive index of the target water body and combined with Snell's law, the divergence angle of the laser beam in the water is determined; Based on the divergence angle, the relationship between the radius of the light spot in the water and the depth is determined, and the geometric flattening factor is determined based on the relationship.

[0011] Optionally, the step of normalizing the net photon count based on the exposure time, the geometry / system gain, the voxel volume, and the geometry flattening factor to obtain the normalized count density for each voxel includes: For each voxel, the exposure time, the geometry / system gain, and the voxel volume are multiplied sequentially to obtain a product, and the net photon count is divided by the product to obtain the volume-normalized count result. The normalized count result after volume normalization is multiplied by the geometric flattening factor to obtain the normalized count density.

[0012] Optionally, determining the diffuse attenuation coefficient field of the downstream irradiance of the target water body based on the normalized count density includes: The beam attenuation coefficient corresponding to each voxel is determined based on the normalized count density of each voxel; The beam attenuation coefficient is mapped to the diffuse attenuation coefficient field of the downstream irradiance of the target water body by engineering proportion.

[0013] Optionally, determining the beam attenuation coefficient corresponding to each voxel based on the normalized count density of each voxel includes: By applying mild second-order regular smoothing, vertical smoothing is performed based on the vertical sequence characteristics of the normalized count density to obtain a smoothed count sequence. Perform a natural logarithmic operation on the smoothed counting sequence to obtain a logarithmic counting profile. The depth slope of the logarithmic counting profile was determined by locally weighted linear regression. The depth slope is converted into the beam attenuation coefficient of the voxel using the volume scattering lidar equation.

[0014] Optionally, mapping the beam attenuation coefficient to the downstream diffuse attenuation coefficient field of the target water body through engineering scaling includes: Based on the preset reference points for the diffuse attenuation coefficient of downlink irradiance, the ratio coefficient between the beam attenuation coefficient and the diffuse attenuation coefficient of downlink irradiance is determined by the least squares fitting method. Based on the scaling factor, the beam attenuation coefficient is mapped to the downward irradiance diffuse attenuation coefficient corresponding to the voxel; A downward irradiance diffuse decay coefficient field is generated based on the downward irradiance diffuse decay coefficient corresponding to each voxel.

[0015] Secondly, the downlink irradiance diffuse attenuation coefficient extraction system of the present invention includes: The data acquisition module is used to perform a circular scan of the target water body using a photon counting lidar to obtain photon point cloud data of the target water body; The depth conversion module is used to determine the depth data corresponding to each photon based on the elevation data of each photon in the photon point cloud data, and generate depth-converted photon point cloud data based on the depth data. An exposure time calculation module is used to determine the exposure time of the horizontal pixel column corresponding to each photon based on the geometric characteristics of the circular scan. The voxelization and background removal module is used to perform three-dimensional voxelization on the depth-converted photon point cloud data to obtain a three-dimensional voxel grid of the depth-converted photon point cloud data, and to perform counting and background removal processing on the three-dimensional voxel grid to obtain the voxel volume and net photon count of each voxel in the three-dimensional voxel grid. The calculation module is used to determine the geometric / system gain of each voxel through a geometric / system gain model, and to determine the geometric flattening factor according to Snell's law and a preset water refractive index. The density generation module is used to normalize the net photon count based on the exposure time, the geometry / system gain, the voxel volume, and the geometry flattening factor to obtain the normalized count density of each voxel. The attenuation coefficient inversion module is used to determine the diffuse attenuation coefficient field of the downstream irradiance of the target water body based on the normalized count density.

[0016] The method and system for extracting the diffuse attenuation coefficient of downlink irradiance of this invention firstly involves a circular scan of the target water body using a photon-counting lidar to acquire photon point cloud data, thus ensuring that the photon point cloud data contains rich water body information. Next, depth data is determined based on the elevation data of each photon, generating depth-converted photon point cloud data. This depth conversion allows subsequent data processing to be performed with the water surface as a reference, eliminating errors caused by elevation differences. Furthermore, the exposure time of the horizontal pixel column corresponding to each photon is determined based on the geometric characteristics of the circular scan. In actual measurements, due to the specific scanning method, photons at different locations may have different exposure times, directly affecting the accuracy of photon counting. This invention, by calculating the exposure time, can correct the photon count under different exposure times, thereby more accurately reflecting the true condition of the water body. Subsequently, the depth-converted photon point cloud data is voxelized into a 3D voxel mesh, and then subjected to counting and background removal processing to count the number of photons within each voxel. Simultaneously, background removal processing eliminates interference from non-target signals, such as solar background and dark counts, effectively reducing the impact of noise on the measurement results and improving data purity and reliability. The geometric / system gain of each voxel is determined using a geometric / system gain model, and the geometric flattening factor is determined based on Snell's law and a preset water refractive index. The geometric / system gain model considers the characteristics of the lidar system itself, such as the number of emitted photons or the optical efficiency of the receiver, as well as the optical properties of the water, such as the transmittance of the air / water interface. The geometric flattening factor is used to correct for geometric effects caused by the divergence of the laser beam in the water. Through the geometric / system gain and geometric flattening factor, this invention can more accurately describe the photon propagation process in water, thereby more precisely extracting the downward irradiance diffuse attenuation coefficient. Finally, the net photon count is normalized based on exposure time, geometric / system gain, voxel volume, and geometric flattening factor to obtain the normalized count density for each voxel. By comprehensively considering all factors, the influence of differences in exposure time, geometry, and system characteristics between different voxels is eliminated. The normalized count density is controlled only by the optical properties of the water body, namely, volumetric scattering and beam attenuation, thus more accurately reflecting the optical properties of the water body. The downlink irradiance diffuse attenuation coefficient field of the target water body is accurately determined based on the normalized count density. In summary, this invention significantly improves the accuracy of obtaining the downlink irradiance diffuse attenuation coefficient through depth conversion, exposure time correction, three-dimensional voxelization and background removal, geometric / system gain modeling, and normalization, providing a more accurate basis for water transparency, euphotic depth, and water quality ecological assessment. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a method for extracting the diffuse attenuation coefficient of downlink irradiance in one embodiment of the present invention. Figure 2 This is a schematic diagram of a three-dimensional point cloud in one embodiment of the present invention; Figure 3 This is a schematic diagram of spatial grid blocks in one embodiment of the present invention; Figure 4 This is a top-view schematic diagram of water surface scanning in one embodiment of the present invention; Figure 5 This is a schematic diagram of the process for extracting the diffuse attenuation coefficient of downlink irradiance in another embodiment of the present invention. Detailed Implementation

[0018] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Although some embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the present invention. It should be understood that the accompanying drawings and embodiments of the present invention are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0019] It should be understood that the various steps described in the method embodiments of the present invention may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0020] The term "comprising" and its variations as used herein are open-ended, meaning "including but not limited to"; the term "based on" means "at least partially based on"; the term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments"; and the term "optionally" means "optional embodiments". Definitions of other terms will be given in the following description. It should be noted that the concepts of "first," "second," etc., mentioned in this invention are used only to distinguish different devices, modules, or units, and are not intended to limit the order of functions performed by these devices, modules, or units or their interdependencies.

[0021] It should be noted that the terms "one" and "more" used in this invention are illustrative rather than restrictive. Those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0022] The names of the messages or information exchanged between the multiple devices in the embodiments of the present invention are for illustrative purposes only and are not intended to limit the scope of these messages or information.

[0023] Combination Figure 1As shown in the figure, an embodiment of the present invention provides a method for extracting the diffuse attenuation coefficient of downlink irradiance, comprising: The target water body is scanned in a circular manner using a photon-counting lidar to obtain photon point cloud data of the target water body.

[0024] Specifically, photon-counting lidar, as an advanced detection device, can detect extremely weak light signals with high sensitivity. It performs a circular scan of the target water body, which comprehensively covers the water area and acquires rich water information. The resulting photon point cloud data contains a large amount of discrete photon information, with each photon having its corresponding coordinates and other data. This step utilizes photon-counting lidar to provide high-quality raw data. The photon point cloud data can be represented as a three-dimensional point cloud, such as... Figure 2 As shown, the original 3D point cloud of photons includes surface point clouds and underwater point clouds. The formulaic representation of photon point cloud data is as follows: ; in, It is the horizontal coordinate (m) of the i-th photon. Elevation (m) For timestamps (s). N It represents the total number of photon points collected. For photon point cloud data sets, For set The i-th element in.

[0025] Based on the elevation data of each photon in the photon point cloud data, the depth data corresponding to each photon is determined, and based on the depth data, depth-converted photon point cloud data is generated.

[0026] Specifically, by analyzing the elevation data of each photon in the photon point cloud data and converting it into depth data, the positional information of the photons can be transformed from an elevation dimension to a depth dimension, which better meets the needs of water body optical property research. Generating depth-converted photon point cloud data allows subsequent processing and calculations to be performed with the water surface as a reference, providing an accurate depth reference for accurately extracting the diffuse attenuation coefficient of downflow irradiance and helping to improve extraction accuracy.

[0027] Based on the geometric characteristics of the circular scan, the exposure time of the horizontal pixel column corresponding to each photon is determined.

[0028] Specifically, during circular scanning, the exposure time of the corresponding horizontal pixel pillar may differ for photons at different positions due to factors such as scanning trajectory and speed. Exposure time directly affects the photon counting result; failure to correct the exposure time can lead to deviations in photon counting, thus affecting the extraction of the downlink irradiance attenuation coefficient. Therefore, this invention considers the geometric characteristics of circular scanning to determine the exposure time of the horizontal pixel pillar corresponding to each photon, providing a basis for subsequent photon counting correction, thereby improving data accuracy and reducing systematic errors caused by inconsistent exposure times.

[0029] The depth-converted photon point cloud data is subjected to 3D voxelization to obtain a 3D voxel grid of the depth-converted photon point cloud data. Counting and background removal are then performed on the 3D voxel grid to obtain the voxel volume and net photon count of each voxel in the 3D voxel grid.

[0030] Specifically, 3D voxelization is a method of discretizing spatial data, which can divide complex photon point cloud data into regular voxel grids. This allows for convenient statistical analysis of photons within each voxel. During the counting process, the number of photons within each voxel is counted, and the voxel volume is recorded, providing data support for subsequent normalization processing. Background removal is to remove interference from non-target signals, such as solar background light and dark counts. These interference signals increase data noise and reduce the accuracy of downlink irradiance diffuse attenuation coefficient extraction. Through background removal, a purer photon count, i.e., a net photon count, can be obtained, thereby improving data quality and laying the foundation for subsequent accurate calculations.

[0031] The geometric / system gain of each voxel is determined by a geometric / system gain model, and the geometric flattening factor is determined according to Snell's law and the preset water refractive index.

[0032] Specifically, the geometric / system gain model comprehensively considers the characteristics of the lidar system, such as the number of emitted photons, receiving optical efficiency, and detection quantum efficiency, as well as the optical properties of the water body, such as the transmittance of the air / water interface. Determining the geometric / system gain of each voxel through the geometric / system gain model allows for a more accurate description of the photon propagation process in the water. Furthermore, Snell's law, the fundamental law describing the refraction of light in different media, combined with a preset water refractive index, allows for the determination of the geometric flattening factor. The geometric flattening factor is used to correct for the geometric effects caused by the divergence of the laser beam in the water, avoiding misinterpreting this geometric divergence as optical attenuation in the water. This results in a more accurate reflection of the true optical properties of the water body and improves the accuracy of extracting the diffuse attenuation coefficient of the downlink irradiance.

[0033] The net photon count is normalized based on the exposure time, the geometry / system gain, the voxel volume, and the geometry flattening factor to obtain the normalized count density for each voxel.

[0034] Specifically, normalization is the process of standardizing the net photon counts of different voxels. By considering factors such as exposure time, geometry / system gain, voxel volume, and geometry flattening factor, the influence of differences in these factors between different voxels can be eliminated. The normalized count density is only controlled by the optical properties of the water body, namely, volumetric scattering and beam attenuation, thus more accurately reflecting the optical properties of the water body and obtaining a normalized count density that can be directly used to calculate the diffuse attenuation coefficient of downlink irradiance.

[0035] The diffuse attenuation coefficient field of the downstream irradiance of the target water body is determined based on the normalized count density.

[0036] Specifically, the normalized count density contains important information about the optical properties of water bodies. By analyzing and calculating the normalized count density, the downflow irradiance diffuse attenuation coefficient field of the target water body is obtained. The downflow irradiance diffuse attenuation coefficient field is the distribution of the downflow irradiance diffuse attenuation coefficient with spatial location (such as horizontal position and depth) within a specific spatial region. The downflow irradiance diffuse attenuation coefficient is a key parameter for measuring the ability of a water body to attenuate visible light, and its distribution can reflect important information such as the transparency of the water body, the depth of the euphotic zone, and the ecological status of the water quality.

[0037] The method for extracting the diffuse attenuation coefficient of downlink irradiance in this invention first involves a circular scan of the target water body using a photon-counting lidar to acquire photon point cloud data, thus ensuring the data contains rich information about the water body. Next, depth data is determined based on the elevation data of each photon, generating depth-converted photon point cloud data. This depth conversion allows subsequent data processing to be performed with the water surface as the reference, eliminating errors caused by elevation differences. Furthermore, the exposure time of the corresponding horizontal pixel column for each photon is determined based on the geometric characteristics of the circular scan. In actual measurements, due to the specific scanning method, photons at different locations may have different exposure times, directly affecting the accuracy of photon counting. This invention, by calculating the exposure time, can correct the photon count under different exposure times, thereby more accurately reflecting the true condition of the water body. Subsequently, the depth-converted photon point cloud data is voxelized into a 3D voxel mesh, and then subjected to counting and background removal processing to count the number of photons within each voxel. Simultaneously, background removal processing eliminates interference from non-target signals, such as solar background and dark counts, effectively reducing the impact of noise on the measurement results and improving data purity and reliability. The geometric / system gain of each voxel is determined using a geometric / system gain model, and the geometric flattening factor is determined based on Snell's law and a preset water refractive index. The geometric / system gain model considers the characteristics of the lidar system itself, such as the number of emitted photons or the optical efficiency of the receiver, as well as the optical properties of the water, such as the transmittance of the air / water interface. The geometric flattening factor is used to correct for geometric effects caused by the divergence of the laser beam in the water. Through the geometric / system gain and geometric flattening factor, this invention can more accurately describe the photon propagation process in water, thereby more precisely extracting the downward irradiance diffuse attenuation coefficient. Finally, the net photon count is normalized based on exposure time, geometric / system gain, voxel volume, and geometric flattening factor to obtain the normalized count density for each voxel. By comprehensively considering all factors, the influence of differences in exposure time, geometry, and system characteristics between different voxels is eliminated. The normalized count density is controlled only by the optical properties of the water body, namely, volumetric scattering and beam attenuation, thus more accurately reflecting the optical properties of the water body. The downlink irradiance diffuse attenuation coefficient field of the target water body is accurately determined based on the normalized count density. In summary, this invention significantly improves the accuracy of obtaining the downlink irradiance diffuse attenuation coefficient through depth conversion, exposure time correction, three-dimensional voxelization and background removal, geometric / system gain modeling, and normalization, providing a more accurate basis for water transparency, euphotic depth, and water quality ecological assessment.

[0038] Optionally, the step of determining the depth data corresponding to each photon based on the elevation data of each photon in the photon point cloud data, and generating depth-converted photon point cloud data based on the depth data, includes: The space corresponding to the target water body is divided into multiple spatial grid blocks according to a preset size; By performing equal-width histogram statistics on the elevation data of each photon within the spatial grid block, the mode bin of the elevation distribution of the photon within each spatial grid block is obtained; The water surface elevation of the spatial grid block is determined based on the midpoint of the mode bin of the spatial grid block; Based on the water surface elevation, the elevation data of each photon is converted into the depth data; The elevation data in the photon point cloud data is replaced with the depth data to generate depth-converted photon point cloud data.

[0039] Specifically, firstly, the space corresponding to the target water body is divided into multiple spatial grid blocks of preset size. Then, by performing equal-width histogram statistics on the photon elevation data within each spatial grid block, the mode bin for the photon elevation distribution within each spatial grid block is obtained. The equal-width histogram statistics are an effective data distribution analysis method that can intuitively reflect the distribution of photon elevations in various intervals, while the mode bin represents the area with the highest concentration of photon elevations within that spatial grid block, thus providing a reliable basis for determining the water surface elevation. Determining the water surface elevation based on the midpoint of the mode bin accurately reflects the actual position of the water surface and avoids misjudgments of water surface elevation due to individual abnormal elevation data. Subsequently, using the determined water surface elevation as a benchmark, the elevation data of each photon is converted into depth data. Finally, the elevation data in the photon point cloud data is replaced with depth data to generate depth-converted photon point cloud data.

[0040] Combination Figure 3 As shown, the space is divided into grid blocks according to Δx×Δy. Within each spatial block R, a histogram of equal width is used to statistically analyze the elevation z. , ; in, For the first Photon count per elevation box; For the first The elevation of a photon. As the starting point of the histogram, For the width of the histogram box, It is an integer used to define the number of intervals or the number of iterations of a process.

[0041] Determine the mode bin index Take the midpoint of the box as the water surface elevation. : ; Convert elevation to depth with the water surface as zero, defining downward as positive: ; The distance from each underwater photon to the water surface, i.e., the depth value, is given by... Replace the original .

[0042] In this embodiment of the invention, by dividing spatial grid blocks and performing histogram statistics of equal width, a large amount of discrete photon point cloud data can be effectively processed, accurately determining the water surface elevation and avoiding depth calculation errors caused by inaccurate water surface elevation determination in traditional methods. The photon point cloud data after depth conversion can better reflect the optical characteristic distribution inside the water body, improving the accuracy and reliability of subsequent processing steps, thereby enhancing the accuracy and robustness of the entire downlink irradiance diffuse attenuation coefficient extraction method.

[0043] Optionally, determining the exposure time of the horizontal pixel column corresponding to each photon based on the geometric characteristics of the circular scan includes: Based on the photon point cloud data after depth conversion, the horizontal region of the target water body is divided into multiple horizontal pixel pillars according to a preset pixel size, wherein each horizontal pixel pillar corresponds to multiple photons; Based on the geometric characteristics of the circular scan and the trajectory parameters of the photon counting lidar, the exposure time of each horizontal pixel column is obtained by discrete accumulation calculation.

[0044] Specifically, the horizontal pixel column division is not a simple spatial grid cutting, but rather a process based on the depth-converted photon point cloud data. The target water body's horizontal region is discretized into vertical cylindrical units in three-dimensional space according to a preset pixel size (e.g., Δx×Δy). Since each photon is assigned to a unique horizontal pixel column through its horizontal coordinates, and the laser spot will cover the same horizontal region multiple times during circular scanning, a single horizontal pixel column necessarily corresponds to multiple photons. This multi-photon-single-pixel column correspondence avoids the interference of the statistical randomness of individual photons and improves computational efficiency through batch processing of pixel columns, laying the spatial unit foundation for subsequent batch calculation of exposure time. Discrete accumulation calculation of exposure time is based on the kinematic essence of circular scanning, transforming the abstract scanning process into quantifiable time statistics. Its core is to utilize the geometric characteristics of circular scanning (e.g., half-angle of the scanning cone, radius of the water surface spot) and lidar trajectory parameters (e.g., platform horizontal position, flight altitude, scanning angular velocity, initial azimuth angle) to construct a dynamic trajectory model of the water surface spot center, and then achieve exposure time statistics through time-step discretization. Specifically, the observation period is divided into preset time steps. At each time step, it is determined whether the center of the light spot falls within the coverage area of ​​the current horizontal pixel column. Finally, the duration of all time steps that meet the conditions is accumulated to obtain the total exposure time of the pixel column.

[0045] Combination Figure 4 As shown, the trajectory of the center of the water surface spot in the top view of the water surface scan is: ; Where X and Y represent the horizontal position of the platform, and H represents the flight altitude. To scan the half-angle of the cone, This refers to the scanning angular velocity; This is the initial azimuth angle.

[0046] Pixel bar exposure time (discrete accumulation): ; in, The horizontal pixel center is Δx, and Δy are the pixel dimensions. , Half diagonal of pixels, Let be the radius of the light spot on the water surface, and Δt be the time step, 1( ) is the coverage indicator function. For single-photon event timestamps In time At any given moment, the horizontal position coordinates of the lidar platform.

[0047] In the two formulas above, t is a continuous time variable used to describe the changes in scanning geometry and spot coverage over time, and is related to the single-photon event timestamp t. nDifferent. The exposure time E of each pixel is obtained by discretely sampling the time axis and accumulating the coverage duration. ij This provides a timescale benchmark for subsequent statistical normalization and physical modeling.

[0048] Optionally, the step of performing three-dimensional voxelization on the depth-converted photon point cloud data to obtain a three-dimensional voxel grid of the depth-converted photon point cloud data, and performing counting and background removal processing on the three-dimensional voxel grid to obtain the voxel volume and net photon count of each voxel in the three-dimensional voxel grid, includes: Based on the spatial distribution of the depth-converted photon point cloud data, three-dimensional voxelization is performed to construct the three-dimensional voxel mesh. Obtain the photon count within each voxel falling into the three-dimensional voxel grid and the voxel volume of each voxel; The background count rate of each horizontal pixel bar is determined based on the signal within the pulse time window; Based on the background count rate and the preset effective sampling duration of the voxel layer, the expected background count of the voxels is obtained. The net photon count of each voxel is obtained based on the difference between the photon count of each voxel and the expected background count of the voxel.

[0049] Specifically, constructing a 3D voxel grid based on the spatial distribution of the 3D photon point cloud data after transforming the depth coordinates is not a simple spatial division. Instead, it is based on the horizontal (x, y) and depth (z) coordinate range of the point cloud, determining the grid starting point (x0, y0, z0) and cell boundaries according to a preset resolution (e.g., Δx×Δy×Δz), ensuring that the voxel grid can completely cover the target water area without generating redundant space. This transforms the originally disordered discrete photons into 3D counting units with clear spatial assignments. Secondly, the photon count is obtained by counting the number of photons falling within the boundary of each voxel, directly reflecting the intensity of the laser-water interaction within that spatial unit; the voxel volume is calculated by preset resolution parameters, providing key parameters for subsequent elimination of spatial size differences between different voxels and unifying the counting and statistical benchmark. The off-pulse time window is the period outside the laser emission pulse. During this period, there is no effective signal of laser-water interaction, only noise such as solar background light and device dark current. By counting the number of photons in the horizontal pixel column during this period and dividing by the time length, the background noise intensity of that pixel column can be accurately quantified. Then, the effective sampling time of the voxel layer is related to the laser detection time at the depth of the voxel. By multiplying the background count rate by the effective sampling time, the expected background count of the voxel contributed by noise in each voxel can be obtained.

[0050] Construct a three-dimensional voxel mesh and count the number of photons in each voxel. The voxel volume is recorded to provide stable statistics for subsequent normalization and profile slope.

[0051] The definition of a voxel and its volume representation are as follows: ; Where: Δx×Δy×Δz is the voxel resolution. , , Align with the grid starting point according to the region envelope. ΔV ijk The volume is the voxel volume.

[0052] The voxel count is: ; For voxel counting, i.e., falling into voxels The number of photons in the event. Each point (photon event) is denoted as . .therefore, Let be the spatial coordinates of the l-th photon.

[0053] Estimate non-target signals such as solar background and dark count to obtain background-free counts. for: ; in, The background count rate of the pixel column. The effective sampling duration for the voxel layer. For the expected background count of voxels, This is the net count after background removal.

[0054] In this embodiment of the invention, the net photon count is obtained by the difference between the photon count and the expected background count, which completely eliminates the interference of background noise on the effective signal. The net photon count reflects only the true result of the scattering and attenuation effect of the laser and the water body, thus solving the problem in related technologies where background interference leads to falsely high or low counts, thereby affecting the accuracy of the attenuation coefficient.

[0055] Optionally, determining the geometric / system gain of each voxel using a geometric / system gain model, and determining the geometric flattening factor based on Snell's law and a preset water refractive index, includes: Obtain the system parameters and optical characteristic parameters of the photon counting lidar; Based on the system parameters and the optical characteristic parameters, a geometric / system gain model is constructed, and the geometric / system gain of each voxel is determined through the geometric / system gain model. Based on the preset refractive index of the target water body and combined with Snell's law, the divergence angle of the laser beam in the water is determined; Based on the divergence angle, the relationship between the radius of the light spot in the water and the depth is determined, and the geometric flattening factor is determined based on the relationship.

[0056] Specifically, the system parameters encompass inherent hardware attributes such as the number of emitted photons, receiving optical efficiency, detection quantum efficiency, and electronic gate width. Optical characteristic parameters include dynamic transmission parameters such as the emitter-receiver overlap function, air / water interface transmittance (including the Freund term), and voxel slant distance. These parameters directly determine the degree to which non-aquatic factors affect photon counting during the laser's transmission and reception process, thus avoiding misinterpreting these non-aquatic factors as water attenuation characteristics. Secondly, a geometric / system gain model is constructed based on the system parameters and optical characteristic parameters, and the geometric / system gain of each voxel is determined. This is not a simple parameter superposition but a physical derivation based on the volume scattering lidar equation. Next, the divergence angle of the laser beam in water is determined based on the preset water refractive index and Snell's law, focusing the geometric characteristics of the conical beam for circular scanning. Snell's law establishes a correspondence between the laser propagation angles in air and water, thereby deriving the divergence angle of the laser beam in water. Finally, the relationship between the radius of the light spot in water and the depth is determined based on the divergence angle, and the geometric flattening factor is determined to accurately correct the divergence effect of the conical beam. The radius of the light spot in water increases with depth (determined by the divergence angle), and the geometric flattening factor needs to be constructed based on this relationship to counteract the geometric dilution effect that the deeper the depth, the larger the light spot, and the fewer the counts per unit volume. This ensures that the corrected counts can truly reflect the scattering and attenuation capabilities of the water body itself, rather than the interference of geometric shape changes.

[0057] Specifically, the geometry / system gain is: ; in, The proportionality constant is a system constant that takes into account factors such as the number of emitted photons, receiving optical efficiency, detection quantum efficiency, and electron gate width. The transmit-receive overlap function takes into account the receive field of view (FOV), pointing deviation, and beam-track offset caused by refraction. (Dimensionless) represents the air / water interface permeability (including the Freund term), typically 0.9. 0.98, y is the slant distance; the Euclidean distance from the platform position to the voxel center.

[0058] Conical divergence and flattening are: ; ; ; ; in, The refractive index of water is 1.33, which is commonly used for green light. ,r Let be the propagation angle and divergence angle in water; approximated by Snell's law. The radius of the light spot in the water. The radius at the water surface; This is a flattening factor. The radius of the light spot in the water. Radius of the light spot at the water surface; The angle of incidence is denoted as .

[0059] In this embodiment of the invention, non-aquatic interference from hardware and optical transmission is precisely isolated through voxel-level calculation of geometric / system gain, thereby eliminating non-aquatic factors and allowing the counting results to initially focus on the interaction between water and laser, avoiding counting interpretation deviations caused by system parameter fluctuations and optical path differences. On the other hand, by introducing a geometric flattening factor, the divergence effect of the circular scanning conical beam in water is effectively corrected, significantly reducing the inversion error of the downward irradiance diffuse attenuation coefficient caused by non-aquatic factors.

[0060] Optionally, the step of normalizing the net photon count based on the exposure time, the geometry / system gain, the voxel volume, and the geometry flattening factor to obtain the normalized count density for each voxel includes: For each voxel, the exposure time, the geometry / system gain, and the voxel volume are multiplied sequentially to obtain a product, and the net photon count is divided by the product to obtain the volume-normalized count result. The normalized count result after volume normalization is multiplied by the geometric flattening factor to obtain the normalized count density.

[0061] Specifically, a systematic elimination process is achieved through step-by-step correction and quantitative calculation, ultimately ensuring that the normalized count density is only related to the optical properties of the water body itself (opposite volume scattering coefficient β, beam attenuation coefficient c). First, the net photon count is divided by the product of exposure time, geometry / system gain, and voxel volume to eliminate three key types of interference. Dividing by exposure time resolves the counting bias caused by uneven coverage time in circular scanning, avoiding the misjudgment of high counts due to prolonged scanning as water attenuation. Dividing by geometry / system gain isolates the influence of system hardware (emitted photon count, receiving efficiency, etc.) and optical transmission (interface transmittance, slant range attenuation, etc.), ensuring that the count is independent of interference from equipment and transmission paths. Dividing by voxel volume unifies the spatial scale differences of three-dimensional voxels, eliminating the spatial bias of "high counts for large-volume voxels and low counts for small-volume voxels." The three factors work together to transform the net photon count into a volume-normalized count, initially achieving batch removal of non-water body factors. Secondly, the volume normalization count is multiplied by the geometric flattening factor to correct the conical beam divergence effect of circular scanning. The geometric flattening factor is constructed based on the relationship between the radius of the light spot in water and the depth. The dilution effect can be compensated by multiplication, avoiding misjudging the geometric shape change as a difference in water body attenuation.

[0062] Net count Normalize exposure, gain, and volume, and apply divergent flattening to obtain an image that is only affected by water optics ( , Controlled normalized count density : ; ; in, The exposure time for each pixel column; For geometric / system gain; The volume is the voxel volume. This is a flattening factor. For physical count density, external factors have been stripped away, and only follow... , change; The scattering coefficient is the volumetric scattering coefficient. The beam attenuation coefficient; (Dimensionless) represents the distance from the water surface to the depth. The cumulative attenuation of optical thickness.

[0063] In this embodiment of the invention, by first stripping away the basic interference and then correcting the geometric divergence, an interference elimination closed loop is formed, ensuring that the normalized count density is controlled only by the water body's anti-radial scattering coefficient and beam attenuation coefficient.

[0064] Optionally, determining the diffuse attenuation coefficient field of the downstream irradiance of the target water body based on the normalized count density includes: The beam attenuation coefficient corresponding to each voxel is determined based on the normalized count density of each voxel; The beam attenuation coefficient is mapped to the diffuse attenuation coefficient field of the downstream irradiance of the target water body by engineering proportion.

[0065] Specifically, after constructing the normalized count density, the normalized count density of each voxel has been stripped of external factors such as exposure, geometry, system, and volume, and is only controlled by water body scattering and beam attenuation. Through slight vertical smoothing and logarithmic processing, exponential attenuation is transformed into a near-linear logarithmic profile, which facilitates the robust calculation of the beam attenuation coefficient using the logarithm-depth slope of locally weighted linear regression. Subsequently, the beam attenuation coefficient is mapped to a downlink irradiance diffuse attenuation coefficient field using an engineering scale, thereby realizing the conversion from normalized count density to target downlink irradiance diffuse attenuation coefficient field.

[0066] In this embodiment of the invention, by utilizing the exponential decay characteristics of the volume scattering lidar equation, and through mathematical modeling and calculation, the complex physical process is transformed into operable algorithmic steps, thereby achieving accurate extraction of the diffuse decay coefficient field of downstream irradiance in water.

[0067] Optionally, the step of determining the beam attenuation coefficient corresponding to each voxel based on the normalized count density of each voxel includes: By applying mild second-order regular smoothing, vertical smoothing is performed based on the vertical sequence characteristics of the normalized count density to obtain a smoothed count sequence. Perform a natural logarithmic operation on the smoothed counting sequence to obtain a logarithmic counting profile. The depth slope of the logarithmic counting profile was determined by locally weighted linear regression. The depth slope is converted into the beam attenuation coefficient of the voxel using the volume scattering lidar equation.

[0068] Specifically, to determine the beam attenuation coefficient based on the normalized count density of each voxel, a slight second-order regularized smoothing of the vertical sequence of normalized count density within each horizontal pixel column is first applied. This stabilizes the vertical continuity of the count sequence, reduces the interference of individual voxel count fluctuations on subsequent calculations, and avoids over-smoothing that masks the true vertical variation in water attenuation. Then, the logarithm of the smoothed sequence is taken to transform the exponential attenuation relationship of laser propagation in water into a near-linear logarithmic profile. The depth slope of the logarithmic profile is then calculated using locally weighted linear regression: a depth window is constructed centered on a single voxel, and the count weights of voxels within the window (such as weights positively correlated with the count quantity) are used to solve for the slope of the logarithmic count as a function of depth. Finally, the slope is multiplied by a coefficient -1 / 2 (to compensate for the two-way transmittance effect of the laser's downward and return journeys) to obtain the beam attenuation coefficient corresponding to that voxel.

[0069] Utilizing the exponential decay characteristic of the volume scattering lidar equation, the depth slope of the logarithmic profile (multiplied by...) 1 / 2) Directly converted into beam attenuation coefficient.

[0070] ; ; in, ={k r,…,k+r} is a layer Centered depth window (covering 2 layers) +1), Let u be the depth value of the data point. The window weights (dimensionless) can be Gaussian kernels or weights related to counting. The local slope is logarithmic to depth; Beam attenuation coefficient; coefficient Half comes from two-way transmittance. , , The weighted average of all voxel depth values. This is the weighted mean of the normalized count density of all voxels. For the target water body at location The estimated value of the diffuse attenuation coefficient of the downlink irradiance at that location.

[0071] In this embodiment of the invention, the vertical smoothing and logarithmic processing based on the normalized count density effectively reduces the interference of random fluctuations in the count. The slope solution method of local weighted linear regression not only conforms to the physical nature of the exponential attenuation of laser propagation, but also adapts to the differences in attenuation characteristics of water bodies at different depths, so that the beam attenuation coefficient can truly reflect the attenuation ability of water bodies to lasers.

[0072] Optionally, mapping the beam attenuation coefficient to the downstream diffuse attenuation coefficient field of the target water body through engineering scaling includes: Based on the preset reference points for the diffuse attenuation coefficient of downlink irradiance, the ratio coefficient between the beam attenuation coefficient and the diffuse attenuation coefficient of downlink irradiance is determined by the least squares fitting method. Based on the scaling factor, the beam attenuation coefficient is mapped to the downward irradiance diffuse attenuation coefficient corresponding to the voxel; A downward irradiance diffuse decay coefficient field is generated based on the downward irradiance diffuse decay coefficient corresponding to each voxel.

[0073] Specifically, a quantitative correlation between the two is established through a small number of reference sample points: sample points in the target water body with existing measured downward irradiance diffuse attenuation coefficients are selected, and the beam attenuation coefficients corresponding to the sample points are fitted with the measured values ​​using least squares to calibrate a unified engineering scaling factor; then, using voxels as units, the beam attenuation coefficient of each voxel is multiplied by this scaling factor to obtain the downward irradiance diffuse attenuation coefficient of a single voxel; finally, the coefficients of all voxels are arranged according to their spatial coordinates (horizontal x, y and depth z) to form a three-dimensional downward irradiance diffuse attenuation coefficient field covering the target water body, realizing the engineering transformation from single-point parameters to a three-dimensional field.

[0074] Specifically, Convert to target 3D for: ; ; in, The proportionality constant (dimensionless) is determined by least-squares calibration using a small number of reference samples Ω. For reference, the downward slow decay; This is the result of the inversion. is the beam attenuation coefficient.

[0075] Combination Figure 5 As shown, an embodiment of the present invention provides a downlink irradiance diffuse attenuation coefficient extraction system, comprising: The data acquisition module is used to perform a circular scan of the target water body using a photon counting lidar to obtain photon point cloud data of the target water body; The depth conversion module is used to determine the depth data corresponding to each photon based on the elevation data of each photon in the photon point cloud data, and generate depth-converted photon point cloud data based on the depth data. An exposure time calculation module is used to determine the exposure time of the horizontal pixel column corresponding to each photon based on the geometric characteristics of the circular scan. The voxelization and background removal module is used to perform three-dimensional voxelization on the depth-converted photon point cloud data to obtain a three-dimensional voxel grid of the depth-converted photon point cloud data, and to perform counting and background removal processing on the three-dimensional voxel grid to obtain the voxel volume and net photon count of each voxel in the three-dimensional voxel grid. The calculation module is used to determine the geometric / system gain of each voxel through a geometric / system gain model, and to determine the geometric flattening factor according to Snell's law and a preset water refractive index. The density generation module is used to normalize the net photon count based on the exposure time, the geometry / system gain, the voxel volume, and the geometry flattening factor to obtain the normalized count density of each voxel. The attenuation coefficient inversion module is used to determine the diffuse attenuation coefficient field of the downstream irradiance of the target water body based on the normalized count density.

[0076] The downlink irradiance diffuse attenuation coefficient extraction system of the present invention has the same advantages over the prior art as the above-mentioned downlink irradiance diffuse attenuation coefficient extraction method, and will not be repeated here.

[0077] While the present invention has been disclosed above, its scope of protection is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention, and all such changes and modifications will fall within the scope of protection of the present invention.

Claims

1. A method for extracting a downwelling irradiance diffuse attenuation coefficient, characterized in that, include: The target water body is scanned in a circular manner by a photon counting lidar to obtain the photon point cloud data of the target water body; Based on the elevation data of each photon in the photon point cloud data, the depth data corresponding to each photon is determined, and the depth-converted photon point cloud data is generated based on the depth data. Based on the geometric characteristics of the circular scan, the exposure time of the horizontal pixel column corresponding to each photon is determined; The depth-converted photon point cloud data is subjected to three-dimensional voxelization to obtain a three-dimensional voxel grid of the depth-converted photon point cloud data. Counting and background removal processing are performed on the three-dimensional voxel grid to obtain the voxel volume and net photon count of each voxel in the three-dimensional voxel grid. The geometric / system gain of each voxel is determined by a geometric / system gain model, and the geometric flattening factor is determined according to Snell's law and the preset water refractive index. The net photon count is normalized based on the exposure time, the geometry / system gain, the voxel volume, and the geometry flattening factor to obtain the normalized count density of each voxel. The diffuse attenuation coefficient field of the downstream irradiance of the target water body is determined based on the normalized count density.

2. The method of claim 1, wherein, The step of determining the depth data corresponding to each photon based on the elevation data of each photon in the photon point cloud data, and generating depth-converted photon point cloud data based on the depth data, includes: The space corresponding to the target water body is divided into multiple spatial grid blocks according to a preset size; By performing equal-width histogram statistics on the elevation data of each photon within the spatial grid block, the mode bin of the elevation distribution of the photon within each spatial grid block is obtained; The water surface elevation of the spatial grid block is determined based on the midpoint of the mode bin of the spatial grid block; Based on the water surface elevation, the elevation data of each photon is converted into the depth data; The elevation data in the photon point cloud data is replaced with the depth data to generate depth-converted photon point cloud data.

3. The method of claim 1, wherein, Determining the exposure time of the horizontal pixel column corresponding to each photon based on the geometric characteristics of the circular scan includes: Based on the photon point cloud data after depth conversion, the horizontal region of the target water body is divided into multiple horizontal pixel pillars according to a preset pixel size, wherein each horizontal pixel pillar corresponds to multiple photons; Based on the geometric characteristics of the circular scan and the trajectory parameters of the photon counting lidar, the exposure time of each horizontal pixel column is obtained by discrete accumulation calculation.

4. The method of claim 3, wherein the downwelling irradiance diffuse attenuation coefficient is extracted by, The process of performing 3D voxelization on the depth-converted photon point cloud data to obtain a 3D voxel mesh of the depth-converted photon point cloud data, and performing counting and background removal processing on the 3D voxel mesh to obtain the voxel volume and net photon count of each voxel in the 3D voxel mesh, includes: Based on the spatial distribution of the depth-converted photon point cloud data, three-dimensional voxelization is performed to construct the three-dimensional voxel mesh. Obtain the photon count within each voxel falling into the three-dimensional voxel grid and the voxel volume of each voxel; The background count rate of each horizontal pixel bar is determined based on the signal within the pulse time window; Based on the background count rate and the preset effective sampling duration of the voxel layer, the expected background count of the voxels is obtained. The net photon count of each voxel is obtained based on the difference between the photon count of each voxel and the expected background count of the voxel.

5. The method for extracting the diffuse attenuation coefficient of downlink irradiance according to claim 4, characterized in that, The process of determining the geometric / system gain of each voxel using a geometric / system gain model and determining the geometric flattening factor based on Snell's law and a preset water refractive index includes: Obtain the system parameters and optical characteristic parameters of the photon counting lidar; Based on the system parameters and the optical characteristic parameters, a geometric / system gain model is constructed, and the geometric / system gain of each voxel is determined through the geometric / system gain model. Based on the preset refractive index of the target water body and combined with Snell's law, the divergence angle of the laser beam in the water is determined; Based on the divergence angle, the relationship between the radius of the light spot in the water and the depth is determined, and the geometric flattening factor is determined based on the relationship.

6. The method for extracting the diffuse attenuation coefficient of downlink irradiance according to claim 1, characterized in that, The step of normalizing the net photon count based on the exposure time, the geometry / system gain, the voxel volume, and the geometry flattening factor to obtain the normalized count density for each voxel includes: For each voxel, the exposure time, the geometry / system gain, and the voxel volume are multiplied sequentially to obtain a product, and the net photon count is divided by the product to obtain the volume-normalized count result. The normalized count result after volume normalization is multiplied by the geometric flattening factor to obtain the normalized count density.

7. The method for extracting the diffuse attenuation coefficient of downlink irradiance according to claim 1, characterized in that, The step of determining the diffuse attenuation coefficient field of the downstream irradiance of the target water body based on the normalized count density includes: The beam attenuation coefficient corresponding to each voxel is determined based on the normalized count density of each voxel; The beam attenuation coefficient is mapped to the diffuse attenuation coefficient field of the downstream irradiance of the target water body by engineering proportion.

8. The method for extracting the diffuse attenuation coefficient of downlink irradiance according to claim 7, characterized in that, The step of determining the beam attenuation coefficient corresponding to each voxel based on the normalized count density of each voxel includes: By applying mild second-order regular smoothing, vertical smoothing is performed based on the vertical sequence characteristics of the normalized count density to obtain a smoothed count sequence. Perform a natural logarithmic operation on the smoothed counting sequence to obtain a logarithmic counting profile. The depth slope of the logarithmic counting profile was determined by locally weighted linear regression. The depth slope is converted into the beam attenuation coefficient of the voxel using the volume scattering lidar equation.

9. The method for extracting the diffuse attenuation coefficient of downlink irradiance according to claim 7, characterized in that, The process of mapping the beam attenuation coefficient to the diffuse attenuation coefficient field of the target water body through engineering proportions includes: Based on preset reference points for the diffuse attenuation coefficient of downlink irradiance, the ratio coefficient between the beam attenuation coefficient and the diffuse attenuation coefficient of downlink irradiance is determined by the least squares fitting method. Based on the scaling factor, the beam attenuation coefficient is mapped to the downward irradiance diffuse attenuation coefficient corresponding to the voxel; A downward irradiance diffuse decay coefficient field is generated based on the downward irradiance diffuse decay coefficient corresponding to each voxel.

10. A system for extracting the diffuse attenuation coefficient of downlink irradiance, characterized in that, include: The data acquisition module is used to perform a circular scan of the target water body using a photon counting lidar to obtain photon point cloud data of the target water body; The depth conversion module is used to determine the depth data corresponding to each photon based on the elevation data of each photon in the photon point cloud data, and generate depth-converted photon point cloud data based on the depth data. An exposure time calculation module is used to determine the exposure time of the horizontal pixel column corresponding to each photon based on the geometric characteristics of the circular scan. The voxelization and background removal module is used to perform three-dimensional voxelization on the depth-converted photon point cloud data to obtain a three-dimensional voxel grid of the depth-converted photon point cloud data, and to perform counting and background removal processing on the three-dimensional voxel grid to obtain the voxel volume and net photon count of each voxel in the three-dimensional voxel grid. The calculation module is used to determine the geometric / system gain of each voxel through a geometric / system gain model, and to determine the geometric flattening factor according to Snell's law and a preset water refractive index. The density generation module is used to normalize the net photon count based on the exposure time, the geometry / system gain, the voxel volume, and the geometry flattening factor to obtain the normalized count density of each voxel. The attenuation coefficient inversion module is used to determine the diffuse attenuation coefficient field of the downstream irradiance of the target water body based on the normalized count density.

Images