Hyperspectral water depth inversion method, device, equipment, storage medium and program product

Abstract

The invention provides a hyperspectral water depth inversion method, device and equipment, a storage medium and a program product, and is applied to the technical field of water depth inversion. The method comprises the following steps: acquiring hyperspectral data; based on the Beer-Lambert law, the reflectivity spectral data of the hyperspectral data are converted into absorbance spectral data, radiation transmission model transformation of absorbance is carried out, and the sensitivity degree of the absorbance spectral data to the water depth is larger than that of the reflectivity spectral data to the water depth; establishing a nonlinear equation set of spectral data and water depth parameters according to the radiation transmission model and the absorbance spectral data; and solving the nonlinear equation set to obtain a water depth inversion result.

Description

Technical Field

[0001] The present invention relates to the technical field of water depth inversion, and particularly to a hyperspectral water depth inversion method, device, equipment, storage medium and program product. Background Art

[0002] Shallow water areas such as coastal zones, islands and reefs, and coral reefs provide key environmental elements for the biodiversity and sustainable development of marine ecosystems. By performing water depth inversion on shallow water areas, the obtained water depth data can be applied to fields such as navigation safety, port planning, marine engineering design, marine scientific research, and marine resource development. Therefore, it is crucial to perform water depth inversion on the nearshore and shallow sea areas.

[0003] In the prior art, with the successful launch of many remote sensing satellites, the research on water depth inversion of coastal zones and around islands and reefs using remote sensing technology has also made great progress, forming water depth inversion methods based on various data such as multispectral, synthetic aperture radar, and hyperspectral. Among them, for clear sky multispectral data and all-weather images to invert the water depth of nearshore shallow waters, empirical regression models of single band and band ratio are adopted, and the mapping relationship between water depth and remote sensing data is established through calibration with on-site measured water depth data to achieve water depth inversion of shallow sea areas. Clear sky hyperspectral data can provide rich spectral information of shallow sea areas. At the same time, since the parameters of the water body radiative transfer model contain physical quantities of shallow water depth, the parameters in the water body radiative transfer model are inverted through hyperspectral data, and water depth inversion can be achieved without calibration with on-site measured water depth data.

[0004] However, due to the high absorption coefficient and low reflectivity value of the water body, the response value of water body pixels in hyperspectral images is low and the range is small. Therefore, the weak water body radiation signal is easily submerged in background noise, which is not conducive to water depth inversion of different water body pixels. Summary of the Invention

[0005] The present invention provides a hyperspectral water depth inversion method, device, equipment, storage medium and program product to solve the problem of low water depth inversion accuracy caused by weak water body radiation signals.

[0006] The present invention provides a hyperspectral water depth inversion method, including: acquiring hyperspectral data; converting the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law, and performing a radiative transfer model transformation of absorbance, where the sensitivity of the absorbance spectral data to water depth is greater than that of the reflectance spectral data to water depth; establishing a non-linear equation set of spectral data and water depth parameters according to the radiative transfer model and the absorbance spectral data; and solving the non-linear equation set to obtain the water depth inversion result.

[0007] A hyperspectral water depth inversion method provided by the present invention, converting the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law, includes: According to the formula converting the reflectance spectral data of the hyperspectral data into absorbance spectral data; where A represents absorbance, I represents the intensity of reflected radiation, represents the intensity of incident radiation, and R represents reflectance.

[0008] A hyperspectral water depth inversion method provided by the present invention, obtaining hyperspectral data, includes: collecting original hyperspectral data through a hyperspectral detector; performing preprocessing operations on the original hyperspectral data to obtain surface reflectance data, and the preprocessing operations sequentially include radiometric correction, geometric correction, and atmospheric correction; screening water body pixels for water depth inversion from the surface reflectance data, and correcting abnormal water body pixels caused by solar glint due to water surface undulation to obtain the hyperspectral data.

[0009] A hyperspectral water depth inversion method provided by the present invention, after converting the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law and performing a radiative transfer model transformation of the absorbance, the method further includes: filtering the absorbance spectral data based on a filtering coefficient with a given window size and order.

[0010] A hyperspectral water depth inversion method provided by the present invention, after converting the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law and performing a radiative transfer model transformation of the absorbance, the method further includes: parameterizing the unknown vector parameters in the radiative transfer model equation.

[0011] A hyperspectral water depth inversion method provided by the present invention, solving the nonlinear equations to obtain the water depth inversion result, includes: solving the nonlinear equations through a hybrid optimization algorithm to obtain the water depth inversion result; where the hybrid optimization algorithm is generated based on a particle swarm optimization algorithm and a trust region strategy optimization algorithm.

[0012] The present invention also provides a hyperspectral water depth inversion device, including the following modules: an acquisition module and a processing module; the acquisition module is used to acquire hyperspectral data; the processing module is used to convert the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law, and perform a radiative transfer model transformation of the absorbance, and the sensitivity of the absorbance spectral data to water depth is greater than the sensitivity of the reflectance spectral data to water depth; establishing a nonlinear equation set of spectral data and water depth parameters according to the radiative transfer model and the absorbance spectral data; solving the nonlinear equation set to obtain the water depth inversion result.

[0013] A hyperspectral water depth inversion device provided by the present invention, the processing module is configured to: According to the formula Convert the reflectance spectral data of the hyperspectral data into absorbance spectral data; where A represents absorbance, I represents the intensity of reflected radiation, represents the intensity of incident radiation, and R represents reflectance.

[0014] A hyperspectral water depth inversion device provided by the present invention, the acquisition module is configured to collect original hyperspectral data through a hyperspectral detector; perform preprocessing operations on the original hyperspectral data to obtain surface reflectance data, and the preprocessing operations sequentially include radiometric calibration, geometric calibration, and atmospheric calibration; screen water body pixels for water depth inversion from the surface reflectance data, and correct abnormal water body pixels caused by solar glint due to water surface undulation to obtain the hyperspectral data.

[0015] A hyperspectral water depth inversion device provided by the present invention, after converting the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law and performing a radiative transfer model transformation of the absorbance, the processing module is configured to filter the absorbance spectral data based on filter coefficients with a given window size and order.

[0016] A hyperspectral water depth inversion device provided by the present invention, after converting the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law and performing a radiative transfer model transformation of the absorbance, the processing module is configured to parameterize the unknown vector parameters in the radiative transfer model equation.

[0017] A hyperspectral water depth inversion device provided by the present invention, the processing module is configured to solve the non-linear equations through a hybrid optimization algorithm to obtain a water depth inversion result; wherein, the hybrid optimization algorithm is generated based on a particle swarm optimization algorithm and a trust region strategy optimization algorithm.

[0018] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the program, it implements the hyperspectral water depth inversion method as described in any one of the above.

[0019] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the hyperspectral water depth inversion method as described in any one of the above.

[0020] The present invention also provides a computer program product, including a computer program which, when executed by a processor, implements the hyperspectral water depth inversion method as described in any one of the above.

[0021] The hyperspectral water depth inversion method, device, equipment, storage medium and program product provided by the present invention can convert the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law, and perform a radiative transfer model transformation of the absorbance. Since the absorbance spectral data is more sensitive to water depth than the reflectance spectral data, the weak information of the water body radiation signal can be enhanced, making the spectral of water body pixels at different water depths more distinguishable, thereby improving the accuracy of hyperspectral data water depth inversion. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0023] Figure 1 is a schematic flowchart of the hyperspectral water depth inversion method provided by the present invention; Figure 2 is a flowchart of the hybrid optimization algorithm provided by the present invention; Figure 3 is a schematic structural diagram of the hyperspectral water depth inversion device provided by the present invention; Figure 4 is a schematic structural diagram of the electronic equipment provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0024] To make the objectives, technical solutions and advantages of the present application clearer, the following will clearly and completely describe the technical solutions in the present application with reference to the drawings in the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments in the present application belong to the scope of protection of the present application.

[0025] It should be noted that in the embodiments of the present application, words such as "exemplary" or "for example" are used to represent examples, illustrations or explanations. Any embodiment or design solution described as "exemplary" or "for example" in the embodiments of the present application should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Rather, the use of words such as "exemplary" or "for example" is intended to present relevant concepts in a specific manner.

[0026] It should be noted that in this article, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising that element. In addition, it should be pointed out that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in the reverse order according to the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0027] For the convenience of clearly describing the technical solutions of the embodiments of the present application, in the embodiments of the present application, terms such as "first" and "second" are used to distinguish the same or similar items with basically the same functions and effects. Those skilled in the art can understand that the terms "first", "second", etc. are not intended to limit the quantity and execution order.

[0028] Some exemplary embodiments are described for the purpose of illustration in the embodiments of the present application. It should be understood that the present application can be implemented in other ways not specifically shown in the drawings.

[0029] As Figure 1 shown, the embodiments of the present application provide a hyperspectral water depth inversion method, which can be applied to a hyperspectral water depth inversion device. The hyperspectral water depth inversion method may include S101 - S104: S101. The hyperspectral water depth inversion device acquires hyperspectral data.

[0030] Optionally, the acquiring of the hyperspectral data includes: collecting original hyperspectral data through a hyperspectral detector; performing preprocessing operations on the original hyperspectral data to obtain surface reflectance data, and the preprocessing operations sequentially include radiometric correction, geometric correction, and atmospheric correction; screening water body pixels for water depth inversion from the surface reflectance data, and correcting abnormal water body pixels caused by solar glint due to water surface undulation to obtain the hyperspectral data.

[0031] Specifically, the hyperspectral water depth inversion device can first collect the original hyperspectral data through a hyperspectral detector. The original hyperspectral data is the response value after the photoelectric conversion of the hyperspectral detector, which is usually called the DN (Digital Number) value. This value has no physical meaning and needs to be converted into the corresponding radiance value through relative and absolute radiometric calibration processing. The conversion formula is shown in the following formula (1): ; (1) Where L represents the radiance value, the subscript i represents the band index, and k and b represent the calibration gain and bias after radiometric calibration of the hyperspectral detector, respectively.

[0032] After that, the hyperspectral water depth inversion device can select the matching homologous points of the image to be processed and the reference data based on the reference image data; then, according to the selected matching homologous points, establish a rational polynomial of the hyperspectral image transformation model; then, use the control point image coordinates and geographic coordinates, and adopt the least squares estimation method to solve the rational polynomial coefficients; finally, solve the geographic coordinates of each pixel of the image one by one according to the obtained transformation parameters, and resample and output the original image to obtain a geometrically corrected hyperspectral image with geographic coordinates including the shallow water area.

[0033] Next, the hyperspectral water depth inversion device can perform atmospheric correction on the hyperspectral image data using the atmospheric radiative transfer model. The atmospheric radiative transfer model describes the radiance value entering the sensor entrance pupil as the sum of direct radiation, indirect radiation, and path radiation, as shown in the following formula (2): ; (2) Where is the surface reflectance corresponding to the pixel, is the average surface reflectance of the area around the pixel, is the spherical albedo of the atmosphere, is the radiance of the atmospheric backscattering, and A and B are coefficients depending on the atmospheric and geometric conditions.

[0034] By establishing a lookup table of water vapor content and model parameters, the input hyperspectral radiance data can be converted into surface reflectance data above the surface and water surface.

[0035] After completing the preprocessing operation, the hyperspectral water depth inversion device can screen out the water pixels for water depth inversion from the surface reflectance data and establish a water spectral index (WSI), as shown in the following formula (3): ; (3) Where R is the reflectance, the subscript g is the green light band, and the subscript r is the near-infrared band.

[0036] The hyperspectral water depth inversion device can judge each pixel in the hyperspectral image one by one. When the WSI value is greater than 0, the pixel is a water body pixel and a corresponding image mask is generated.

[0037] Finally, since the waters in the shallow sea area are close to the shore and usually have large wave undulations, the reflected radiation of the water body pixels in the hyperspectral image may be affected by the sun glint caused by the wave undulations, resulting in abnormal remote sensing reflectance of the water body in the image and no longer being suitable for water depth inversion. The hyperspectral water depth inversion device can adopt the water reflectance index model correction method. The water reflectance index model can describe the reflected radiation of the water body as two parts, one part is the reflected radiation of the water body itself, and the other part is the specular reflection generated by the water surface (i.e., sun glint), as shown in the following formula (4): ; (4) Where is the total reflected radiation of the water body, is the reflected radiation of the water body itself, is the sun glint reflectance, is the sun glint radiation.

[0038] The hyperspectral water depth inversion device can calculate using the multi-band brightness threshold method, remove the sun glint reflectance in the reflectance, and obtain the hyperspectral data.

[0039] S102. The hyperspectral water depth inversion device converts the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law and performs a radiation transfer model transformation of the absorbance.

[0040] Among them, the sensitivity of the absorbance spectral data to the water depth is greater than that of the reflectance spectral data to the water depth.

[0041] Optionally, the hyperspectral water depth inversion device converts the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law, including: According to the formula convert the reflectance spectral data of the hyperspectral data into absorbance spectral data; where A represents absorbance, I represents the intensity of reflected radiation, represents the intensity of incident radiation, and R represents reflectance.

[0042] Specifically, the radiant energy reflected by water bodies to hyperspectral detectors is relatively low. Generally, for typical water body pixels in hyperspectral images, the reflectance in the visible light band is lower than 0.2, and the reflectance value in the infrared band is close to 0. In current hyperspectral remote sensing water depth inversion methods, the water surface reflectance is used to invert the water bottom depth. However, the reflectance signal is weak and has a small variation range, which is not conducive to accurate water depth inversion.

[0043] In the Beer-Lambert law, which describes the absorption characteristics of light when propagating in a solution, a new physical quantity - absorbance is proposed. This physical quantity is used to represent the light absorption ability of the solution, and the absorbance is proportional to the absorption coefficient of the solution and the path length of the light passing through. Therefore, for nearshore shallow waters, absorbance is a more suitable physical quantity for water depth inversion than reflectance. The present invention proposes to use a spectral transformation method to transform the reflectance spectral data into absorbance spectral data to enhance the weak spectral information of water bodies. The specific spectral transformation formula is shown in the following formula (5): ; (5) where A represents absorbance, I represents the intensity of reflected radiation, represents the intensity of incident radiation, and R represents reflectance.

[0044] According to the above formula (5), it can be seen that the reflectance data is first inverted and then logarithmically transformed to obtain the absorbance data. After spectral transformation, the numerical value of the spectral data increases, and the range is stretched to [0, 3].

[0045] It should be noted that due to the strong light absorption ability of water bodies, the water body absorbance signal is significantly improved compared to the water body reflectance signal. Therefore, the spectral of water body pixels with different water depths can have stronger distinguishability, thereby improving the accuracy of water depth inversion of hyperspectral data.

[0046] Optionally, after converting the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law, the hyperspectral water depth inversion device can filter the absorbance spectral data based on the filter coefficients with a given window size and order.

[0047] Specifically, by designing more imaging bands or narrower imaging band widths, hyperspectral detectors can obtain spectral data with high spectral resolution. However, the consequent problem is that the radiant energy entering a single imaging band is significantly reduced. At the same time, due to the inaccuracy of the radiometric calibration coefficient and the non-uniformity of atmospheric correction, unnatural spikes and pseudo-absorptions are generated in some bands of the hyperspectral data, which are called spectral "artifacts". This application can use the S-G filter to remove spectral "artifacts". That is, the hyperspectral water depth inversion device can first calculate the filtering coefficients for a given window size and order; then, on the spectral dimension of the absorbance hyperspectral data, slide the S-G filter band by band to the last band to obtain the spectral data after removing "artifacts". In this way, unnatural spikes and pseudo-absorptions on the spectrum can be eliminated, background noise can be reduced, and thus the accuracy of hyperspectral data water depth inversion can be improved.

[0048] Optionally, the hyperspectral water depth inversion device can perform a transformation of the radiative transfer model for absorbance based on the Beer-Lambert law.

[0049] Specifically, the radiative transfer model can be the Hydrolight model, and the Hydrolight model defines the remote sensing reflectance above the water surface as the ratio of the radiance leaving the water to the downward irradiance above the water surface. Based on the divergence and reflection factors at the water-air interface, the remote sensing reflectance above the water surface is converted into the remote sensing reflectance below the water surface using the following formula (6): : ; (6) where represents the water-air divergence factor, represents the internal reflection factor at the water-air interface, represents the remote sensing reflectance below the water surface.

[0050] Based on the Beer-Lambert law, that is, the absorption of light is exponentially related to the path length that the light passes through, a physical model of with parameters such as the scattering coefficient, absorption coefficient, bottom albedo, and bottom depth of the water body components can be established, as shown in the following formula (7): ; (7) where is the optical deep water remote sensing reflectance, represents the optical path correction factor, represents the zenith angle between the propagation path of light in water and the vertical direction, is the optical path extension coefficient for water body scattering, is the optical path extension coefficient for bottom sediment scattering, is the seawater attenuation coefficient, represents the water depth, that is, the vertical distance that light travels from the water surface to the bottom, is the seawater bottom albedo, are the inherent optical property coefficients of seawater. The above coefficients can be calculated using the scattering and absorption coefficients of water components.

[0051] S103, a hyperspectral water depth inversion device establishes a nonlinear equation group of spectral data and water depth parameters according to the radiation transmission model and the absorbance spectral data.

[0052] By transforming the reflectance spectrum data into absorbance spectrum data and performing a radiation transfer model transformation of the absorbance, the above formula (6) can be transformed into the following formula (8). At this point, a nonlinear equation group can be established using the absorbance data and formula (8).

[0053] ; (8) in, Represents remote sensing absorbance.

[0054] Optionally, the hyperspectral water depth inversion device can parameterize unknown vector parameters in the radiation transfer model equation.

[0055] Specifically, in the above-established nonlinear equations, since the scattering and absorption coefficients of water components vary with wavelength, only n nonlinear equations can be established for n independent bands, and the number of unknowns is 4n+1. At this time, the equations are "underdetermined" and other constraints need to be introduced. This application adopts the HOPE parameterization process to parameterize the absorption and scattering coefficients of water components that vary with wavelength into a linear function of a specific band. The specific process is as follows: (a) Phytoplankton pigment absorption coefficient , the absorption coefficient at 440nm is used for fitting, as shown in formula (9). is the absorption coefficient of phytoplankton at 440nm, two empirical coefficients , The fitting results of Lee are adopted.

[0056] ; (9) (b) Absorption coefficient of colored soluble organic matter and debris , its parameterization formula is shown in the following formula (10), where: , is the spectral slope, is the sum of the absorption spectra of organic matter and debris at 440nm, taking the average of the spectral slopes, =0.015nm-1.

[0057] ; (10) (c) Backscattering coefficient of suspended particles , which is defined by Equation (11). Among them, is the backscattering coefficient of suspended particles at 400 nm, and Y is an empirical value.

[0058] ; (11) (d) Bottom albedo , and its parametric formula is shown in the following Equation (12). Among them, is the albedo of the water bottom at 550 nm, is the bottom reflection spectrum normalized by the albedo of the bottom at 550 nm. Common bottom types include sand, aquatic plants, seaweeds, etc.

[0059] ; (12) Thus, the parameterization of the Hydrolight model is completed. The wavelength-dependent , , , unknown coefficients in the nonlinear equations are converted into first-order functions of specific wavelength coefficients. At this time, the unknowns in the equations are respectively .

[0060] S104. The hyperspectral water depth inversion device solves the nonlinear equations to obtain the water depth inversion result.

[0061] Specifically, through the previous steps, a nonlinear equation set between the absorbance (A) spectral data and the scattering and absorption coefficients of water body components ( ), the bottom albedo ( ), and the bottom depth ( ) is established, as shown in the following Equation (13). When the number of bands n participating in the water depth inversion > 5, an "overdetermined" equation set can be obtained (where the pure water absorption coefficient and the backscattering coefficient are both measured results in the laboratory), and this "overdetermined" equation set has a least squares solution.

[0062] ; (13) Optionally, the hyperspectral water depth inversion device can solve the nonlinear equations through a hybrid optimization algorithm to obtain the water depth inversion result; among them, the hybrid optimization algorithm is generated based on the Particle Swarm Optimization (PSO) and the Trust Region Policy Optimization (TRPO).

[0063] Specifically, the hyperspectral water depth inversion device can obtain the least squares solution (i.e., the optimal estimation of model parameters) of this non-linear equation set through a hybrid optimization algorithm. This solution problem has characteristics such as a complex equation structure, a high-dimensional solution space (5 dimensions), and a large search range. PSO has a low dependence on the initial values of unknown parameters and has good global search capabilities, but PSO has weak local optimization capabilities, a slow convergence speed, and unstable optimization results. TRPO has efficient local optimization capabilities and can quickly converge to a local optimal solution, but it has a strong dependence on the initial point. When the initial point is far from the optimal solution, the optimization accuracy is low, the convergence speed is slow, and it is easily trapped in a local optimal solution. The hybrid optimization algorithm can combine the two. First, PSO is used to randomly generate multiple initial solutions, and global search is carried out by sharing search information among the particle swarm. Then, the optimization results obtained by the particle swarm search are used as the initial solutions of TRPO. Finally, TRPO is used for fine local optimization to find the global optimal solution faster and more accurately. The robustness of PSO is used to largely make up for the dependence of the TRPO algorithm on the initial solution during the entire optimization process, and TRPO can further improve the accuracy and stability of the PSO optimization results. The hybrid optimization algorithm combines the advantages of the particle swarm optimization algorithm and the trust region strategy optimization algorithm, has low sensitivity to the initial values of unknown parameters, has good global search capabilities, a fast convergence speed during the optimization process, and stable multiple optimization results.

[0064] As Figure 2 shown, it is the flow chart of the hybrid optimization algorithm. The fitness function adopted during the optimization process is the sum of the squared residuals of the model-simulated spectrum and the measured spectrum . The entire optimization process is divided into two stages: The first stage is the PSO algorithm optimization stage. First, PSO is initialized to obtain the inertia factor, learning factor, particle swarm position, particle swarm velocity, fitness function, and particle swarm fitness value. The initial parameters of PSO are set as shown in Table 1: Table 1

[0065] Then, by minimizing the fitness function, the best positions of individual particles and the group history are found. If the iteration stop criterion is not met, each particle shares its search information to obtain a new search velocity , and its calculation formula is as shown in formula (14): ; (14) where is the inertia factor, is the self-learning factor, is the group learning factor, represents the random number generation function, k is the iteration number, is the current position of the particle, is the historical best position of the particle, is the historical best position of the swarm, and continuously updates the particle position; when the final iteration stops, the historical best position of the particle swarm is output to obtain the PSO optimization result.

[0066] The second stage is the TRPO algorithm optimization stage. First, initialize TRPO, and use a solution close to the global optimum obtained by the PSO algorithm as the initial value of the unknown parameters; then start iterative optimization, calculate the fitness value and the Jacobian matrix, and solve the trust region subproblem, as shown in formula (15): ; (15) where, is the fitness function, is the Jacobian matrix, is the trust region size.

[0067] If the iteration stop criterion is not met, update the trust region size, and calculate the new parameter values from the new trust region step size; finally, when the stop criterion is met, output the parameter optimization result to obtain the least squares solution of the nonlinear equation system. Solve this step for each pixel of the hyperspectral data, and extract the water depth ( ) value, then a water depth inversion map corresponding to the hyperspectral image can be obtained.

[0068] The above hybrid optimization algorithm combines the advantages of the particle swarm optimization algorithm and the trust region strategy optimization algorithm, and is used to find the least squares solution of the equation system established between the hyperspectral data of the water body and the radiative transfer model. In a relatively large solution space range, it can perform global fast search and local fine optimization, and at the same time greatly reduce the dependence on the initial values of the unknown parameters.

[0069] In the embodiments of the present application, the reflectance spectral data of the hyperspectral data can be converted into absorbance spectral data based on the Beer-Lambert law, and the radiative transfer model transformation of the absorbance is performed. Since the absorbance spectral data is more sensitive to the water depth than the reflectance spectral data is to the water depth, the weak information of the water body radiation signal can be enhanced, making the spectral of water body pixels with different water depths have stronger distinguishability, thereby improving the accuracy of water depth inversion of hyperspectral data.

[0070] The above mainly introduces the solution provided by the embodiments of the present application from the perspective of methods. To implement the above functions, it includes the corresponding hardware structures and / or software modules for executing each function. Those skilled in the art should easily realize that, in combination with the units and algorithm steps of each example described in the embodiments disclosed in this article, the embodiments of the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a certain function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0071] For the hyperspectral water depth inversion method provided by the embodiments of the present application, the execution subject can be a hyperspectral water depth inversion device, or a control module for hyperspectral water depth inversion in the hyperspectral water depth inversion device. In the embodiments of the present application, taking the hyperspectral water depth inversion device to execute the hyperspectral water depth inversion method as an example, the hyperspectral water depth inversion device provided by the embodiments of the present application is described.

[0072] It should be noted that the embodiments of the present application can divide the functional modules of the hyperspectral water depth inversion device according to the above method examples. For example, each functional module can be divided corresponding to each function, or two or more functions can be integrated into one processing module. The above integrated modules can be implemented in the form of hardware or in the form of software functional modules. Optionally, the division of modules in the embodiments of the present application is illustrative, and is only a logical function division. There may be other division methods in actual implementation.

[0073] As Figure 3 shown, the embodiments of the present application provide a hyperspectral water depth inversion device 300. The hyperspectral water depth inversion device 300 includes: an acquisition module 301 and a processing module 302; the acquisition module 301 is used to acquire hyperspectral data; the processing module 302 is used to convert the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law, and perform a radiative transfer model transformation of the absorbance, and the sensitivity of the absorbance spectral data to water depth is greater than the sensitivity of the reflectance spectral data to water depth; establish a non-linear equation set of spectral data and water depth parameters according to the radiative transfer model and the absorbance spectral data; solve the non-linear equation set to obtain a water depth inversion result.

[0074] Optionally, the processing module 302 is used to: According to the formula Convert the reflectance spectral data of the hyperspectral data into absorbance spectral data; where A represents absorbance, I represents reflected radiation intensity, I represents the incident radiation intensity, and R represents the reflectivity.

[0075] Optionally, the acquisition module 301 is configured to collect original hyperspectral data through a hyperspectral detector; perform preprocessing operations on the original hyperspectral data to obtain surface reflectivity data, and the preprocessing operations sequentially include radiometric calibration, geometric calibration, and atmospheric calibration; screen water body pixels for water depth inversion from the surface reflectivity data, and correct abnormal water body pixels caused by solar glint due to water surface undulation to obtain the hyperspectral data.

[0076] Optionally, after converting the reflectivity spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law and performing a radiative transfer model transformation of the absorbance, the processing module 302 is configured to filter the absorbance spectral data based on a filtering coefficient with a given window size and order.

[0077] Optionally, after converting the reflectivity spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law and performing a radiative transfer model transformation of the absorbance, the processing module 302 is configured to parameterize the unknown vector parameters in the radiative transfer model equation.

[0078] Optionally, the processing module 302 is configured to solve the non-linear equations through a hybrid optimization algorithm to obtain a water depth inversion result; wherein, the hybrid optimization algorithm is generated based on a particle swarm optimization algorithm and a trust region strategy optimization algorithm.

[0079] In the embodiments of the present application, the reflectivity spectral data of the hyperspectral data can be converted into absorbance spectral data based on the Beer-Lambert law and a radiative transfer model transformation of the absorbance is performed. Since the absorbance spectral data is more sensitive to water depth than the reflectivity spectral data, the weak information of the water body radiation signal can be enhanced, making the spectral of water body pixels at different water depths more distinguishable, thereby improving the accuracy of water depth inversion of hyperspectral data.

[0080] Figure 4 Illustrates a schematic physical structure diagram of an electronic device, such as Figure 4As shown, the electronic device may include: a processor 410, a communications interface 420, a memory 430, and a communication bus 440. Among them, the processor 410, the communications interface 420, and the memory 430 communicate with each other through the communication bus 440. The processor 410 can call the logical instructions in the memory 430 to execute the hyperspectral water depth inversion method, which includes: acquiring hyperspectral data; converting the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law, and performing a radiative transfer model transformation of the absorbance. The sensitivity of the absorbance spectral data to water depth is greater than that of the reflectance spectral data to water depth; establishing a non-linear equation set of spectral data and water depth parameters according to the radiative transfer model and the absorbance spectral data; and solving the non-linear equation set to obtain the water depth inversion result.

[0081] In addition, when the logical instructions in the above-mentioned memory 430 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0082] On the other hand, the present invention also provides a computer program product. The computer program product includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the hyperspectral water depth inversion method provided by the above-mentioned various methods. The method includes: acquiring hyperspectral data; converting the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law, and performing a radiative transfer model transformation of the absorbance. The sensitivity of the absorbance spectral data to water depth is greater than that of the reflectance spectral data to water depth; establishing a non-linear equation set of spectral data and water depth parameters according to the radiative transfer model and the absorbance spectral data; and solving the non-linear equation set to obtain the water depth inversion result.

[0083] In another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the hyperspectral water depth inversion method provided by the above-mentioned various methods. The method includes: obtaining hyperspectral data; converting the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law, and performing a radiative transfer model transformation of the absorbance. The sensitivity of the absorbance spectral data to water depth is greater than the sensitivity of the reflectance spectral data to water depth; establishing a non-linear equation set between the spectral data and the water depth parameter according to the radiative transfer model and the absorbance spectral data; and solving the non-linear equation set to obtain the water depth inversion result.

[0084] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative work.

[0085] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the above technical solutions, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disc, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0086] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A hyperspectral water depth inversion method, characterized in that, Including: Obtaining hyperspectral data; Converting the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law, and performing a radiative transfer model transformation of the absorbance, wherein the sensitivity of the absorbance spectral data to water depth is greater than that of the reflectance spectral data to water depth; Establishing a non-linear equation set of spectral data and water depth parameters according to the radiative transfer model and the absorbance spectral data; Solving the non-linear equation set to obtain a water depth inversion result.

2. The hyperspectral water depth inversion method according to claim 1, wherein The converting the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law includes: According to the formula Convert the reflectance spectral data of the hyperspectral data into absorbance spectral data; where A represents absorbance, I represents the intensity of reflected radiation, represents the intensity of incident radiation, and R represents reflectance.

3. The hyperspectral water depth inversion method according to claim 1, wherein The obtaining the hyperspectral data includes: Collecting original hyperspectral data through a hyperspectral detector; Performing preprocessing operations on the original hyperspectral data to obtain surface reflectance data, and the preprocessing operations sequentially include radiometric correction, geometric correction, and atmospheric correction; Screening water body pixels for water depth inversion from the surface reflectance data, and correcting abnormal water body pixels caused by solar glint due to water surface undulation to obtain the hyperspectral data.

4. The hyperspectral water depth inversion method according to claim 1, wherein After the converting the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law and performing a radiative transfer model transformation of the absorbance, the method further includes: Filtering the absorbance spectral data based on a filtering coefficient with a given window size and order.

5. The hyperspectral water depth inversion method according to claim 1, wherein After the converting the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law and performing a radiative transfer model transformation of the absorbance, the method further includes: Performing parameterization processing on unknown vector parameters in the radiative transfer model equation.

6. The hyperspectral water depth inversion method according to any one of claims 1-5, characterized in that The solving the non-linear equation set to obtain a water depth inversion result includes: Solving the non-linear equation set through a hybrid optimization algorithm to obtain a water depth inversion result; wherein, the hybrid optimization algorithm is generated based on a particle swarm optimization algorithm and a trust region strategy optimization algorithm.

7. A hyperspectral water depth inversion device, characterized in that, Including: An acquisition module and a processing module; The acquisition module is used to obtain hyperspectral data; The processing module is used to convert the reflectance spectral data of the hyperspectral data into absorbance spectral data based on the Beer-Lambert law, and perform a radiative transfer model transformation of the absorbance, wherein the sensitivity of the absorbance spectral data to water depth is greater than that of the reflectance spectral data to water depth; Establishing a non-linear equation set of spectral data and water depth parameters according to the radiative transfer model and the absorbance spectral data; solving the non-linear equation set to obtain a water depth inversion result.

8. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the hyperspectral water depth inversion method according to any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the hyperspectral water depth inversion method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the hyperspectral water depth inversion method according to any one of claims 1 to 6.

Images