In-situ measurement method and system for optical parameters of water body

Abstract

The application discloses an in-situ measurement method and system for optical parameters of a water body, and belongs to the technical field of underwater optical measurement. The method comprises the following steps: synchronously collecting underwater images by using a measurement device composed of at least two imaging units which are not collinear and coplanar and whose relative poses are determined; selecting any imaging unit image as a reference to obtain a plurality of homonymic points with non-reference imaging unit images; for each group of homonymic points, calculating the imaging distance difference between the reference and non-reference units; taking the brightness value of the reference unit image as a reference, calculating the corresponding theoretical brightness value on the non-reference imaging unit image based on the underwater imaging physical model and the initial optical parameters of the water body; calculating the brightness residual by using the calculated theoretical brightness value and the actual imaging observation value; establishing a target optimization function by minimizing the sum of squares of all homonymic point brightness residuals, introducing the red channel absorption characteristics as a regularization condition to constrain the solution space range, and iteratively inverting to obtain the optimal optical parameters of the water body.

Description

Technical Field

[0001] This application relates to the field of underwater optical measurement technology, and in particular to an in-situ measurement method and system for optical parameters in water. Background Technology

[0002] In ocean observation and underwater operations, the absorption, forward scattering, and backscattering of light by water bodies cause image brightness degradation, reduced sharpness, color shift, and decreased contrast. To accurately understand the underwater imaging degradation process and quantitatively restore underwater images, it is necessary to obtain the inherent optical parameters of the water body.

[0003] Traditional underwater optical systems typically focus primarily on image acquisition and visual enhancement, while lacking the ability to measure water optical parameters in real-time, in-situ, and from the same source. This is especially true in environments such as turbid waters, complex near-shore waters, and long-term deep-sea platforms, where water conditions vary significantly over time and space. Offline or intermittent measurements struggle to maintain consistency with the actual imaging time, field of view, and imaging path, resulting in a lack of reliable physical parameter support for image reconstruction and optical sensing results.

[0004] Therefore, how to achieve in-situ, synchronous, and interpretable measurement of optical parameters of water bodies is a technical problem that urgently needs to be solved in related fields. Summary of the Invention

[0005] This application provides a method and system for in-situ measurement of optical parameters of water bodies, which can realize in-situ, synchronous, and interpretable measurement of optical parameters of water bodies.

[0006] To solve the above-mentioned technical problems, this application is implemented as follows: In a first aspect, a method for in-situ measurement of optical parameters in water is provided. The method includes: synchronously acquiring underwater images using a measurement device composed of at least two non-collinear and non-coplanar imaging units determined by their relative poses, wherein the imaging distances from the multiple imaging units to the same point in the underwater region are different; dividing the multiple imaging units into a reference imaging unit and at least one non-reference imaging unit, and using the image acquired by the reference imaging unit as the reference image and the image acquired by the non-reference imaging unit as the non-reference image; acquiring multiple sets of corresponding points between the reference image and at least one of the non-reference images, wherein each set of corresponding points includes the coordinates and brightness observations of a target point in the reference image, and the coordinates and brightness observations of a corresponding point of the target point in at least one of the non-reference images; and calculating the... The imaging distance difference from the target point to each of the corresponding points is used as the reference brightness value. Based on the underwater imaging physical model, the initial water optical parameters, and each imaging distance difference, the theoretical brightness value of each of the corresponding points in the group is calculated. For each corresponding point, the brightness residual is calculated based on the theoretical brightness value and the brightness observation value of the corresponding point in the non-reference image. A target optimization function is established to minimize the sum of squares of the brightness residuals of all corresponding points, and the target optimization function is iteratively solved to obtain the target water optical parameters. The target optimization function uses the red channel absorption characteristics as a regularization term to constrain the solution space range. The target water optical parameters include at least one of the following: water absorption coefficient, water scattering coefficient, water backscattering coefficient, background light intensity, and forward scattering blur parameter.

[0007] Secondly, an in-situ measurement system for optical parameters of a body of water is provided, comprising: an imaging unit module, wherein the imaging unit module is a measurement device composed of at least two non-collinear and non-coplanar imaging units with relative poses determined, and the imaging distances from multiple imaging units to the same point in the underwater region are different; a synchronous acquisition control module, used to control the multiple imaging units to synchronously acquire underwater images; a corresponding point registration module, used to divide the multiple imaging units into a reference imaging unit and at least one non-reference imaging unit, and to use the image acquired by the reference imaging unit as the reference image and the image acquired by the non-reference imaging unit as the non-reference image; acquiring multiple sets of corresponding points between the reference image and at least one non-reference image, wherein each set of corresponding points includes the coordinates and brightness observation value of a target point in the reference image, and the coordinates and brightness observation value of the corresponding point of the target point in at least one non-reference image; and brightness residual. A construction module is used to calculate the imaging distance difference from the target point to each of the corresponding points in each group, using the observed brightness value of the target point as a reference brightness value, and calculating the theoretical brightness value of each of the corresponding points in the group based on the underwater imaging physical model, initial water optical parameters, and each imaging distance difference; and calculating the brightness residual for each corresponding point based on the theoretical brightness value and the observed brightness value of the corresponding point in the non-reference image; a parameter inversion module is used to establish a target optimization function by minimizing the sum of squares of the brightness residuals of all the corresponding points, and iteratively solve the target optimization function to obtain the target water optical parameters, wherein the target optimization function uses the red channel absorption characteristics as a regularization term to constrain the solution space range, and the target water optical parameters include at least one of the following: water absorption coefficient, water scattering coefficient, water backscattering coefficient, background light intensity, and forward scattering blur parameter.

[0008] Thirdly, an electronic device is provided, including a processor and a memory, wherein the memory stores a program or instructions executable on the processor, the program or instructions, when executed by the processor, perform the steps of the method described in the first aspect.

[0009] Fourthly, a readable storage medium is provided, on which a program or instructions are stored, which, when executed by a processor, implement the steps of the method described in the first aspect.

[0010] Fifthly, a computer program product is provided, the computer program product comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions that, when executed by a computer, cause the computer to perform the steps of the method described in the first aspect.

[0011] In this embodiment, an underwater image is synchronously acquired using a measurement device composed of at least two non-collinear and non-coplanar imaging units whose relative poses are determined. The imaging distances from multiple imaging units to the same point in the underwater region are different. The multiple imaging units are divided into a reference imaging unit and at least one non-reference imaging unit, with the image acquired by the reference imaging unit serving as the reference image and the image acquired by the non-reference imaging unit serving as the non-reference image. Multiple sets of corresponding points are obtained between the reference image and at least one non-reference image. Each set of corresponding points includes the coordinates and brightness observations of a target point in the reference image, and the coordinates and brightness observations of the corresponding points of the target point in at least one non-reference image. For each set of corresponding points, the imaging distance difference from the target point to each corresponding point is calculated. The brightness observation of the target point is used as a reference brightness value, and the results are then calculated based on the underwater imaging physical model and the initial water body light... Based on the imaging parameters and the distance difference between each imaging unit, the theoretical brightness value of each corresponding point in the set is calculated. For each corresponding point, the brightness residual is calculated based on the theoretical brightness value and the observed brightness value of the corresponding point in the non-reference image. An objective optimization function is established to minimize the sum of squares of the brightness residuals of all corresponding points, and the objective optimization function is iteratively solved to obtain the target water body optical parameters. The objective optimization function uses the red channel absorption characteristics as a regularization term to constrain the solution space range. The target water body optical parameters include at least one of the following: water body absorption coefficient, water body scattering coefficient, water body backscattering coefficient, background light intensity, and forward scattering blur parameter. This realizes brightness differential observation based on the imaging distance difference of multiple imaging units, which can more realistically reflect the water body optical state during the image formation process, thereby obtaining high-precision and robust water body optical parameter inversion results without relying on high-precision optical instruments.

[0012] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0013] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0014] Figure 1 This paper illustrates a flowchart of an in-situ measurement method for optical parameters of water bodies provided in some embodiments of this application. Figure 2 This illustration shows a schematic diagram of the imaging distance corresponding to multiple imaging units provided in some embodiments of this application; Figure 3 A schematic diagram illustrating the imaging distance difference between multiple imaging units provided in some embodiments of this application is shown; Figure 4A schematic diagram illustrating the observation principle based on the imaging distance difference of dual imaging units provided in some embodiments of this application is shown; Figure 5 A schematic diagram of the parameter inversion process provided in some embodiments of this application is shown; Figure 6 The diagram illustrates a radiation conformance calibration process provided in some embodiments of this application. Figure 7 This paper illustrates another flowchart of an in-situ measurement method for optical parameters of water bodies provided in some embodiments of this application; Figure 8 This paper illustrates another schematic flowchart of an in-situ measurement method for optical parameters of water bodies provided in some embodiments of this application; Figure 9 This invention provides a schematic diagram of the structure of an in-situ measurement system for optical parameters of water bodies according to some embodiments of this application. Figure 10 This paper shows another structural schematic diagram of the in-situ measurement system for optical parameters of water provided in some embodiments of this application; Figure 11 The diagram shows a schematic representation of the structure of an electronic device provided in some embodiments of this application. Detailed Implementation

[0015] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0016] Figure 1 This illustration shows a flowchart of an in-situ measurement method for optical parameters of water bodies provided in an exemplary embodiment of this application. The method can be performed by an electronic device, which may include a terminal and / or a network-side device. The method may include the following steps: S110: A measuring device consisting of at least two non-collinear and non-coplanar imaging units whose relative poses are determined to simultaneously acquire underwater images.

[0017] Among them, the imaging distances from multiple imaging units to the same point in the underwater region are different.

[0018] It is understandable that fixed relative pose refers to a fixed relative positional and orientational relationship between at least two imaging units. Imaging distance refers to the actual distance of light propagation from the target point to the optical center of the imaging unit. The different imaging distances from each imaging unit to the underwater region mean that, for the same point within that underwater region, the spatial positions of different imaging units are pre-differentiated during shooting, resulting in unequal light propagation path lengths from each imaging unit to that point. In other words, each imaging unit acquires images of the same underwater region along different light propagation paths, thus forming multi-view images. For example, as... Figure 2 As shown, there is an observation point in the underwater region, and three imaging units (first imaging unit, second imaging unit, and third imaging unit) observe the observation point from different positions. Because each imaging unit is spatially positioned relative to the observation point, their imaging distances to the observation point are also different: the third imaging unit is closest to the observation point, with the smallest imaging distance; the first imaging unit is next, with a middle imaging distance; and the second imaging unit is farthest from the observation point, with the largest imaging distance.

[0019] In some embodiments, the imaging unit may be an underwater camera, an optical imaging module, a camera with a filter, a color camera, a monochrome camera, a multispectral camera, a hyperspectral camera, or other imaging units.

[0020] In some embodiments, multiple imaging units can be arranged back-to-back along the target observation direction, or arranged at a certain angle or in an array. It should be noted that the multiple imaging units do not overlap in three-dimensional space.

[0021] In some embodiments, synchronous or quasi-synchronous acquisition of multiple imaging units can be achieved through hardware triggering, synchronization controllers, timestamp alignment, or software synchronization. The purpose of synchronization is to ensure that the underwater areas acquired by different imaging units are in approximately the same water and target state, reducing errors introduced by target movement, water disturbance, or lighting changes. For dynamic water bodies or moving platforms, observation data within a short time window can be processed, and data that does not meet the synchronization conditions can be discarded based on timestamp differences.

[0022] S120: Divide the plurality of imaging units into a reference imaging unit and at least one non-reference imaging unit, and use the image acquired by the reference imaging unit as the reference image and the image acquired by the non-reference imaging unit as the non-reference image.

[0023] The reference imaging unit can be the imaging unit closest to the underwater area being measured, the imaging unit furthest from the underwater area being measured, or any imaging unit.

[0024] S130: Obtain multiple sets of corresponding points between the reference image and at least one of the non-reference images.

[0025] Each set of corresponding points includes the coordinates and brightness observations of a target point in the reference image, as well as the coordinates and brightness observations of a corresponding point of the target point in at least one of the non-reference images.

[0026] In some embodiments, acquiring multiple sets of corresponding points between the reference image and at least one non-reference image includes: geometrically calibrating the multiple imaging units by combining surface calibration extrinsic parameters and underwater calibration intrinsic parameters to obtain geometric calibration information, wherein the geometric calibration information includes the intrinsic parameters and relative pose relationships of each imaging unit; based on the geometric calibration information, establishing multiple sets of corresponding points between the reference image and at least one non-reference image by performing distortion correction, field-of-view overlap region determination, and pixel mapping on the reference image and the non-reference image. Geometric calibration can be performed underwater using a calibration board, or pre-calibrated in air followed by underwater refraction correction. Calibration methods can include checkerboard calibration, dot array calibration, global bundle adjustment calibration, etc. It is understood that registration refers to the operation of processing multiple images to establish spatial correspondences between them. Through registration, the same target point in different images is associated, thus forming corresponding points. To improve the stability of parameter estimation, overlapping field-of-view regions can be selected as effective observation areas. Target points within the effective observation area can be used to construct brightness difference observations. That is, through registration processing, for spatial location points appearing in the reference image, corresponding points of the same name are obtained in the non-reference images; spatial location points that appear only in the reference image but not in any non-reference image are discarded.

[0027] In some embodiments, registration may also be achieved by means of image feature matching, camera calibration parameters, affine transformation, homography transformation, epipolar constraints, depth-assisted estimation, or calibration targets.

[0028] In some embodiments, in underwater environments with low visibility or weak target texture, a geometric mapping method based on calibration information can be used to avoid matching failures caused by water turbidity, image blurring, or insufficient texture.

[0029] S140: For each group of corresponding points, calculate the imaging distance difference from the target point to each corresponding point, use the observed brightness value of the target point as the reference brightness value, and calculate the theoretical brightness value of each corresponding point in the group based on the underwater imaging physical model, the initial water optical parameters and each imaging distance difference; and for each corresponding point, calculate the brightness residual based on the theoretical brightness value and the observed brightness value of the corresponding point in the non-reference image.

[0030] It is understood that the imaging distance difference refers to the difference in the light propagation path length from the underwater target point to different imaging units. For example, as... Figure 3 As shown, assuming the first imaging unit is the reference imaging unit, and the second and third imaging units are non-reference imaging units, the imaging distance difference between the first and second imaging units is Δd1, and the imaging distance difference between the first and third imaging units is Δd2. It should be noted that Δd1 and Δd2 are not the actual physical distance between the two imaging units; they are used to represent the difference in imaging distance between the two units. When multiple imaging units observe the same target point in the water from different directions, the brightness of that point recorded by each imaging unit depends on the length of the actual light path from the target point to the imaging unit, i.e., the imaging distance. In the underwater environment, light not only attenuates naturally due to diffusion but is also absorbed and scattered by the water. Energy density decreases with propagation distance; therefore, the more the propagation distance increases, the more light energy is lost. The spatial positions of different imaging units determine their respective optical paths, resulting in differences in the distance light travels from the target point to each imaging unit. This difference will result in varying light energy densities reaching different imaging units, ultimately manifesting as brightness differences of the target point at different viewing angles: the target point appears darker at longer viewing distances and brighter at shorter viewing distances. Therefore, based on brightness observations from a reference image, combined with the imaging distance difference between the non-reference image and the reference image, the theoretical brightness value of the non-reference image is derived using an underwater imaging physical model. This is then compared with the actual observations of the non-reference image to obtain the brightness residual. This brightness residual only includes unknown water optical parameters; therefore, the smaller the residual, the closer the water optical parameters are to their true values.

[0031] In some embodiments, the underwater imaging physical model can be represented as: , Indicates the difference in imaging distance. This indicates the preset optical parameters of the water body. This represents the actual observed brightness value. The target point is acquired using the reference imaging unit. Brightness observations For reference, based on the underwater imaging physical model and preset water body optical parameters Predict target point The theoretical brightness value of the corresponding point at the imaging distance corresponding to the non-reference imaging unit. Then, the brightness values ​​actually observed by the non-reference imaging unit are... With the theoretical brightness value By comparison, the brightness residual is obtained. The brightness residual contains information about absorption, scattering, backscattering, and background light variations caused by the imaging distance difference.

[0032] In this embodiment, the theoretical brightness value is predicted using the imaging distance difference, and the difference between the theoretical value and the actual observation is used as the observation for water body optical parameter inversion. This method can reduce the influence of inherent radiance of unknown scenes, target material differences, and absolute distance uncertainties on parameter inversion, making the brightness observation more controlled by water body optical parameters and imaging distance difference.

[0033] In some embodiments, the difference in imaging distances can be obtained through at least one of the following methods: (1) Calculate based on camera calibration parameters and binocular geometry; (2) Approximate calculation based on the preset baseline length and camera installation direction; (3) Determined with the assistance of structured light, acoustic ranging, laser ranging, or depth sensors; (4) Estimate the target parallax based on the spatial arrangement of the multi-camera array; (5) The calibration is obtained by calibrating a known target or a target at a known distance.

[0034] In some embodiments, the imaging distance difference between a reference imaging unit and a non-reference imaging unit is... Intrinsic parameters (camera focal length) that can be calibrated by two imaging units Image principal point coordinates ), relative extrinsic parameters (lateral baseline length) and longitudinal baseline length The coordinates of the target point corresponding to the reference imaging unit in the two imaging units. The coordinates of corresponding points of non-reference imaging units in the two imaging units. By derivation, for example, the calculation formula can be: ,in, This is the function for calculating the imaging distance difference. This imaging distance difference, as a controllable quantity obtained through pre-design and online calculation of the system, is used to enhance the understanding of the influence of water body optical parameters on changes in observed brightness, transforming parameter inversion from single-frame estimation into an incremental observation problem based on the imaging distance difference.

[0035] In some embodiments, the image can be a grayscale image, a color image, a multispectral image, or a hyperspectral image. For a color image, the brightness residual can be processed separately in the red, green, and blue channels; for a multispectral or hyperspectral image, brightness difference observations can be established separately in multiple bands when determining the brightness residual.

[0036] S150: Establish an objective optimization function by minimizing the sum of squares of the brightness residuals of all the corresponding points, and iteratively solve the objective optimization function to obtain the optical parameters of the target water body.

[0037] The objective optimization function uses the red channel absorption characteristics as a regularization term to constrain the solution space range, and the target water body optical parameters include at least the water body absorption coefficient. Water scattering coefficient Water backscattering coefficient Background light intensity Forward scattering ambiguity parameters One of them.

[0038] In some embodiments, the optical parameters of the target water body may further include: optical attenuation parameters in different bands or different color channels, and equivalent optical parameters related to water turbidity, suspended particulate matter concentration, or water color.

[0039] It is understandable that the brightness residuals corresponding to each of the corresponding points in each group are summarized to establish a target optimization function. For example, this target optimization function... It can be represented as:

[0040] in, The optical parameters of the target water body to be solved are: This represents the total number of objective points that can be used to construct the objective optimization function. Let x be the x-coordinate of the i-th target point. Let be the ordinate of the i-th target point. This represents the brightness residual of the corresponding point at the i-th target point. This brightness residual can be obtained based on the brightness residuals of each corresponding point at the target point, such as the average or maximum value. The objective optimization function is used to optimize the residual between the theoretically predicted brightness and the actual observed brightness. To minimize this, physical constraints, parameter range constraints, smoothing constraints, band correlation constraints, or robust loss terms can be added to the objective function. For example, the water absorption coefficient can be constrained based on the physical properties of the water body. and water scattering coefficient It is a non-negative value and falls within a certain range, such as , ,in, Water absorption coefficient The lower threshold, Water absorption coefficient The upper limit threshold, Water scattering coefficient The lower threshold, Water scattering coefficient The upper limit threshold can constrain the absorption of the red band to be stronger than that of other bands, such as... and ,in, Water absorption coefficient The absorption intensity of R in the red light band, Water absorption coefficient The absorption intensity of G in the green light band, Water absorption coefficient The absorption intensity in the blue light band B can constrain the smooth variation of background light in local areas; a robust loss function can be used to reduce the impact of abnormal pixels, local highly reflective targets, or registration errors on the results.

[0041] In some embodiments, such as Figure 4 The diagram illustrates an observation principle based on the imaging distance difference of two imaging units. This observation process separates the effects of forward scattering and backscattering in the water by inverting the optical parameters of the water body. Specifically, the observation point (imaging distance) ) Measured near-frame observation brightness The brightness of the near-frame observation is mainly composed of backscattering from the water and path attenuation; the brightness of the far observation point (imaging distance) is mainly composed of backscattering from the water and path attenuation. ) Measured brightness of distant frames Due to the longer propagation path, water attenuation and multiple scattering effects are more significant. By comparison... and By combining the differences with the initial water body optical parameters, the target water body optical parameters can be decoupled.

[0042] In some embodiments, the inversion process is as follows: Figure 5 As shown, a target optimization function is first constructed based on pixel-level brightness residuals (i.e., the difference between actual observed brightness and theoretical brightness). The optical parameters of the water body are inverted by minimizing the sum of squares of the brightness residuals of multiple corresponding points. Then, prior knowledge is used to guide the selection of initial values, and the parameters are gradually adjusted through iterative optimization. At the same time, abnormal observation points caused by saturation, excessive darkness, reflection, or registration errors are eliminated. To ensure the physical rationality of the inversion, physical regularization constraints can be embedded, including limiting the parameter range of strong absorption channels, maintaining the correlation between parameters of multiple bands, and applying spatial smoothing constraints. Finally, the target water body optical parameters are output, and an overall evaluation is given based on the parameter confidence interval and the reliability based on residual statistics.

[0043] In some embodiments, during the solution process, an iterative optimization method can be used to solve the objective optimization function and obtain the optical parameters of the target water body. The iterative optimization method can be least squares, weighted least squares, gradient descent, Newton's method, Gauss-Newton method, Levenberg-Marquardt method, Bayesian estimation, particle filtering, Kalman filtering, genetic algorithm, particle swarm optimization, or neural network-assisted optimization.

[0044] In other embodiments, gradient descent can be used to gradually approximate the optimal parameter solution that minimizes the objective function, and the iterative update formula can be: , For the number of iterations, For the first The parameter estimates for the next iteration. For learning rate, For the objective function L, gradient at, For the first The parameter estimates are obtained from +1 iterations. This method balances convergence speed and stability. For multi-camera or multi-baseline scenarios, the brightness residuals corresponding to multiple imaging distance differences can be incorporated into the objective optimization function, thereby improving the robustness of parameter estimation.

[0045] In addition, to avoid local optima, multiple initial value optimization, parameter range constraints, or prior model constraints can be used. To evaluate parameter confidence, parameter confidence can be calculated based on residual distribution, Jacobian matrix, information matrix, repeated observations, or multi-baseline consistency.

[0046] In some embodiments, quantitative information of parameters corresponding to the optical parameters of the target water body can also be output, including: (1) Parameter confidence interval, wherein the parameter confidence interval is used to represent the range of values ​​and confidence level of the parameter; Understandably, the parameter confidence interval is used to represent the range of parameter values ​​and the confidence level. The calculation logic is to determine the interval range by taking the optimal estimated value as the center and combining it with the parameter fluctuation range. It is calculated using the formula: Confidence interval = Parameter estimated value ± 1.96 × Parameter standard deviation. It is used to characterize that the parameter has a 95% probability of falling within the interval. The narrower the interval, the higher the parameter accuracy.

[0047] (2) Parameter reliability, wherein the parameter reliability is used to characterize the accuracy and reliability of the parameter; Understandably, parameter reliability is used to characterize the accuracy and reliability of inversion parameters. Its calculation logic is based on the stability of the residual between theoretical brightness and actual observed brightness. The smaller and more stable the residual, the higher the parameter reliability.

[0048] (3) Whether the imaging distance difference is within the effective observation range, wherein the effective observation range is determined by the upper and lower limits of the imaging distance difference. The lower limit requires that the brightness change caused by the imaging distance difference is greater than the sum of the imaging unit noise and the ambient stray light noise; the upper limit requires that the imaging distance difference will not cause the image acquired by the imaging unit to be submerged by water scattering noise. Understandably, the lower limit ensures that brightness differences can be effectively observed, while the upper limit ensures that distant images have valid signals.

[0049] (4) The distribution of luminance residuals, wherein the distribution of residuals is obtained by statistical calculation of the luminance residuals of all corresponding points; Understandably, the residual values ​​between the actual observed brightness and the theoretically predicted brightness are first calculated point by point. Then, the mean, standard deviation, maximum, minimum, and the percentage of samples whose residuals fall within a reasonable range are calculated for all residuals. This describes the central tendency, dispersion, and distribution of the residuals, which is used to judge the fitting effect between the observed data and the physical model.

[0050] (5) Whether the image acquisition process meets the requirements of radiometric and geometric consistency; (6) Whether there is at least one abnormal state in the image acquisition process, such as saturation, excessive darkness, strong reflection, registration failure, or rapid changes in water body; (7) The optical state level or environmental quality index of the underwater area.

[0051] Understandably, the optical state level of water bodies and environmental quality indicators are calculated and determined based on parameters such as absorption coefficient, scattering coefficient, backscattering coefficient, and attenuation coefficient obtained through inversion. First, the values ​​of attenuation coefficient and backscattering coefficient are used to classify the water body into five levels: clear, relatively clear, moderately turbid, turbid, and extremely turbid. Then, the visibility of the water body is obtained by converting the attenuation coefficient. Combined with the scattering coefficient, the turbidity level of suspended particulate matter is determined. The water color type is judged based on the attenuation ratio of different wavebands. At the same time, the image recovery level is given by comprehensively considering the reliability of parameters and optical level. Finally, a quantitative indicator that can directly reflect the quality of the underwater optical environment is formed.

[0052] In this embodiment, an underwater image is synchronously acquired using a measurement device composed of at least two non-collinear and non-coplanar imaging units whose relative poses are determined. The imaging distances from multiple imaging units to the same point in the underwater region are different. The multiple imaging units are divided into a reference imaging unit and at least one non-reference imaging unit, with the image acquired by the reference imaging unit serving as the reference image and the image acquired by the non-reference imaging unit serving as the non-reference image. Multiple sets of corresponding points are obtained between the reference image and at least one non-reference image. Each set of corresponding points includes the coordinates and brightness observations of a target point in the reference image, and the coordinates and brightness observations of the corresponding points of the target point in at least one non-reference image. For each set of corresponding points, the imaging distance difference from the target point to each corresponding point is calculated. The brightness observation of the target point is used as a reference brightness value, and the results are then calculated based on the underwater imaging physical model and the initial water body light... Based on the imaging parameters and the distance difference between each imaging unit, the theoretical brightness value of each corresponding point in the set is calculated. For each corresponding point, the brightness residual is calculated based on the theoretical brightness value and the observed brightness value of the corresponding point in the non-reference image. A target optimization function is established to minimize the sum of squares of the brightness residuals of all corresponding points, and the target optimization function is iteratively inverted to obtain the target water body optical parameters. The target optimization function uses the red channel absorption characteristics as a regularization term to constrain the solution space range. The target water body optical parameters include at least one of the following: water body absorption coefficient, water body scattering coefficient, water body backscattering coefficient, background light intensity, and forward scattering blur parameter. This realizes brightness differential observation based on the imaging distance difference of multiple imaging units, which can more realistically reflect the water body optical state during the image formation process, thereby obtaining high-precision and robust water body optical parameter inversion results without relying on high-precision optical instruments.

[0053] In some embodiments, the objective optimization function uses the red channel absorption characteristics as a regularization term to constrain the solution space range. This includes: introducing a physical law regularization term during the iterative inversion process, using the non-negativity and value range of the red light band's strong absorption and weak scattering characteristics as constraints, while simultaneously using the wavelength dependence of water absorption and scattering intensity to constrain the blue-green channel parameter range; and incorporating the gradient of the regularization term into the parameter update during each iteration, so that the parameters converge toward the optimal solution fitted to the data while being guided by the physical constraints of the red channel absorption characteristics. Based on the above physical laws, a regularization penalty term is introduced into the original objective optimization function, with no penalty when the water parameters are within a reasonable range, and penalties applied according to the degree of deviation when they exceed the range.

[0054] In some embodiments, before synchronously acquiring underwater images using a measuring device composed of at least two non-collinear and non-coplanar imaging units determined by relative pose, the method further includes: acquiring standard optical test card images at different underwater distances using the measuring device; fitting the radiation deviation between the non-reference imaging unit and the reference imaging unit one by one using the standard optical test card images at multiple distances to obtain a radiation response correction coefficient, a brightness compensation factor, and a color correction matrix; and performing radiation consistency calibration on the plurality of imaging units using the radiation response correction coefficient, the brightness compensation factor, and the color correction matrix.

[0055] It is understandable that, due to potential differences in response, exposure, gain, lens transmittance, or sensor drift between different imaging units, radiometric consistency calibration is required to address the brightness observation differences among multiple imaging units. Radiometric consistency calibration can be performed using a grayscale plate, a standard reflector, a uniform target area, a stable region in a natural scene, or pre-calibrated data. Calibration methods can employ linear response mapping, polynomial mapping, lookup table mapping, channel-specific calibration, or time-window-based dynamic calibration. In some embodiments, by photographing a standard color chart and a grayscale test chart at multiple distances in an experimental water environment, the radiometric deviation between the non-reference imaging unit and the reference imaging unit is fitted one by one, and the radiometric response correction coefficient, brightness compensation factor, and color correction matrix are solved and applied. For example, as... Figure 6 The diagram illustrates the radiometric consistency calibration process. Original images acquired by different imaging units are processed using a grayscale plate or scene constraints to estimate the response mapping relationship, resulting in calibrated images with comparable brightness. After radiometric consistency calibration, the brightness values ​​of different imaging units are comparable, ensuring that subsequent brightness differences or residuals primarily reflect optical attenuation and scattering differences caused by variations in water propagation distance, rather than differences in the response of the imaging units themselves.

[0056] In some embodiments, after establishing a target optimization function to minimize the sum of squares of the brightness residuals of all the corresponding points and iteratively solving the target optimization function to obtain the target water body optical parameters, the method further includes: for each image, inputting the target water body optical parameters into the underwater imaging physical model to obtain the imaging distance output by the underwater imaging physical model, and performing the following steps based on the target water body optical parameters and the imaging distance to reconstruct the image: subtracting the backscattering component from the image based on the target water body optical parameters and the imaging distance to obtain a corresponding first processed image; converting the first processed image to the frequency domain, and... After eliminating the forward scattering component in the frequency domain, the image is inverted back to the spatial domain to obtain a pure direct attenuation component with the forward scattering component removed. The pure direct attenuation component is multiplied by an attenuation compensation factor and color restored to obtain the restored second processed image. Based on the channel features of the blue and green channels, the geometric structure information of the blue and green channels is used as a ground truth reference to repair the edge and texture structure information of the second processed image in the red channel, resulting in a third processed image. Based on the channel features of the red channel, the red channel is used as a scattering ground truth reference to correct the scattering noise of the third processed image in the blue and green channels, resulting in a fourth processed image.

[0057] Understandably, after obtaining the optical parameters of the target water body, these parameters can be incorporated into the underwater imaging physical model to calculate the imaging distance of the underwater target point, and then the underwater image can be physically consistently restored. The restoration process may include at least one of the following: background light estimation, backscattering stripping, forward scattering deblurring, attenuation compensation, and color correction. For example: by subtracting the backscattering component caused by suspended particles from the original image using water optical parameters and imaging distance, a first processed image with background haze noise removed is obtained. Then, the first processed image is transformed to the frequency domain using Fourier transform to eliminate convolutional blur caused by forward scattering. After inverse transform back to the spatial domain, the pure direct attenuation component with only energy attenuation but no scattering interference is extracted. Finally, this component is multiplied by an attenuation compensation factor to compensate for light energy loss. Combined with color restoration to correct color cast caused by wavelength selective absorption, a clear and color-accurate original scene radiation is restored, resulting in a second processed image. The second processed image is then split into RGB three channels. Relying on the characteristics of lower light absorption and higher scene texture fidelity of the blue and green channels, the original structural information such as edges and textures of the image is constrained and repaired to obtain a third processed image. Combining the characteristics of the red channel with the least scattering noise and a clean imaging substrate, the red channel is used as a scattering truth reference to adaptively correct stray scattering noise left in the blue and green channels, accurately eliminating residual scattering interference across the entire domain, and finally outputting a fourth processed image with complete structure and thorough descattering.

[0058] The following is through Figure 7 The illustrated embodiment provides an exemplary description of each process in the above-described in-situ measurement method for optical parameters of water bodies. Figure 7 As shown, the following steps may be included: S701: Construct a measurement device consisting of at least two non-collinear and non-coplanar imaging units whose relative poses are determined.

[0059] At least two imaging units are set up on the underwater observation platform. The relative positional and attitude relationships between the at least two imaging units are known. Multiple imaging units can be arranged forward and backward along the target observation direction, or arranged at a certain angle or in an array. Different imaging units have different imaging distances to the same underwater target point. This results in an imaging distance difference. .

[0060] In other embodiments, three or more imaging units can be set to form multiple imaging distance difference combinations to improve the stability and redundancy of parameter inversion.

[0061] S702: Performs geometric calibration and synchronous acquisition control for at least two imaging units.

[0062] Geometric calibration is performed on each imaging unit to obtain its intrinsic, extrinsic, distortion parameters, and relative pose relationships. Each imaging unit achieves synchronous or quasi-synchronous acquisition through hardware triggering, a synchronization controller, timestamp alignment, or software synchronization. The purpose of synchronization is to ensure that the target scene acquired by different imaging units is in approximately the same water and target state, reducing errors introduced by target movement, water disturbance, or lighting changes. For dynamic water bodies or moving platforms, observation data within a short time window can be processed, and data that does not meet the synchronization conditions can be discarded based on timestamp differences.

[0063] S703: Acquire multi-view underwater images of the same target area.

[0064] Using at least two imaging units , Simultaneous or near-synchronous acquisition of images of the same underwater target area, obtaining at least one close-up image. and at least one distant image . and Based on underwater imaging physical model Establish the degradation process of underwater targets, i.e. , . and The imaging distance from the underwater target to the camera. The physical model for underwater imaging can be the Jaffe-McGlamery model or any of its variants. To preset the optical parameters of the water body, This represents the final required radiance of the underwater target.

[0065] In addition, to improve the stability of parameter estimation, overlapping field-of-view regions can be selected as effective observation regions. Pixels or region blocks within the effective observation region are used to construct brightness difference observations.

[0066] S704: Perform image registration and mapping of corresponding points.

[0067] Image registration can yield multiple sets of corresponding points. Each set of corresponding points corresponds to the brightness observations of the same underwater target point at different propagation distances.

[0068] S705: Determine the imaging distance difference.

[0069] Based on the relative poses between imaging units, camera intrinsic parameters, image coordinates of corresponding points, target geometric relationships, or auxiliary ranging information, determine the imaging distance difference of the same target point in different imaging units. .

[0070] S706: Perform radiation conformity calibration.

[0071] Because there may be differences in response, exposure, gain, lens transmittance, or sensor drift between different imaging units, it is necessary to perform radiometric consistency calibration on the brightness observations of multiple imaging units.

[0072] S707: Construct brightness residual observations based on imaging distance differences.

[0073] This method can reduce the impact of inherent radiance of unknown scenes, differences in target materials, and absolute distance uncertainties on parameter inversion, making the brightness observation more controlled by the optical parameters of the water body and the imaging distance difference.

[0074] S708: Establish the objective optimization function.

[0075] The objective function is used to minimize the sum of squared residuals between the theoretically predicted brightness and the actual observed brightness.

[0076] S709: Solve the objective function.

[0077] The objective optimization function can be solved by iterative optimization methods to obtain the optical parameters of the target water body.

[0078] S710: Outputs optical parameters and quantitative information of the target water body.

[0079] This quantitative information may include at least one of the following: (1) Parameter confidence interval or confidence level. Parameter confidence level is used to characterize the accuracy and reliability of inverted parameters. Its calculation logic is based on the stability of the residual between theoretical brightness and actual observed brightness. The smaller and more stable the residual, the higher the parameter confidence level. Parameter confidence interval is used to represent the range and confidence level of parameter values. The calculation logic is to determine the interval range based on the optimal estimated value and the parameter fluctuation range. It is calculated using the formula: Confidence interval = Parameter estimated value ± 1.96 × Parameter standard deviation. It is used to characterize that the parameter has a 95% probability of falling within this interval. The narrower the interval, the higher the parameter accuracy. (2) Whether the current imaging distance difference is within the effective observation range. The effective observation range is determined by the upper and lower limits of the imaging distance difference. The lower limit requires that the brightness change caused by the imaging distance difference is greater than the sum of the sensor noise and the ambient stray light noise, so as to ensure that the brightness difference can be effectively observed; the upper limit requires that the imaging distance difference will not cause the image acquired by the second imaging unit to be submerged by water scattering noise, so as to ensure that the long-distance image has an effective signal. (3) Residual distribution. The residual distribution is obtained by statistically calculating the brightness residuals of all valid corresponding points. First, the residual values ​​between the actual observed brightness and the theoretical predicted brightness are calculated point by point. Then, the mean, standard deviation, maximum value, minimum value and the proportion of samples with residuals falling within a reasonable range are calculated for all residuals. This describes the central tendency, dispersion and distribution of the residuals, and is used to judge the fitting effect between the observed data and the physical model. (4) Whether the measurements meet the requirements for radiometric and geometric consistency; (5) Whether there are abnormal conditions such as saturation, excessive darkness, strong reflection, registration failure, or rapid changes in the water body; (6) Corresponding water body optical state level or environmental quality index. The water body optical state level and environmental quality index are calculated and determined based on parameters such as absorption coefficient, scattering coefficient, backscattering coefficient and attenuation coefficient obtained by inversion. First, the values ​​of attenuation coefficient and backscattering coefficient are used to classify the water body into five levels: clear, relatively clear, moderately turbid, turbid and extremely turbid. Then, the water body visibility is obtained by converting the attenuation coefficient. The turbidity level of suspended particulate matter is determined by combining the scattering coefficient. The water color type is judged according to the attenuation ratio of different bands. At the same time, the image recovery level is given by combining the reliability of parameters and optical level. Finally, a quantitative index that can directly reflect the underwater optical environment quality is formed.

[0080] S711: Perform physical consistency restoration of underwater images.

[0081] After obtaining the optical parameters of the target water body, these parameters can be substituted into the underwater imaging physical model to calculate the imaging distance of the underwater target point. Then, a physically consistent restoration of the underwater image can be performed. The restoration process may include background light estimation, backscattering stripping, forward scattering deblurring, attenuation compensation, and color correction.

[0082] For example, based on Figure 7 As shown in the embodiment, this application also provides a flowchart of an in-situ measurement method for optical parameters of water bodies, as follows: Figure 8 As shown, the in-situ measurement process of the optical parameters of the water body involves image acquisition, radiometric consistency calibration, geometric registration and acquisition of corresponding points, calculation of imaging distance difference, construction of brightness residuals and parameter inversion, and output of water body optical parameters and quantitative indicators. This process enables the in-situ measurement of water body optical parameters and can further provide physically consistent parameter support for underwater image restoration, target detection and environmental monitoring.

[0083] Figure 9 This illustration shows a schematic diagram of an in-situ measurement system for optical parameters of water bodies according to an embodiment of this application. Figure 10 This application provides an embodiment of an in-situ measurement system for optical parameters of water bodies, as shown in the schematic diagram below. Figure 9 and Figure 10 As shown, the in-situ measurement system 900 for optical parameters of water bodies may include: an imaging unit module 910, a synchronous acquisition and control module 920, a registration module for corresponding points 930, a brightness residual construction module 940, and a parameter inversion module 950.

[0084] In this embodiment, the imaging unit module 910 is a measuring device composed of at least two non-collinear and non-coplanar imaging units with relative poses determined, and the imaging distances from multiple imaging units to the same point in the underwater region are different; the synchronous acquisition control module 920 is used to control the multiple imaging units to synchronously acquire underwater images; the corresponding point registration module 930 is used to divide the multiple imaging units into a reference imaging unit and at least one non-reference imaging unit, and use the image acquired by the reference imaging unit as the reference image and the image acquired by the non-reference imaging unit as the non-reference image; multiple sets of corresponding points are obtained between the reference image and at least one non-reference image, wherein each set of corresponding points includes the coordinates and brightness observation value of a target point in the reference image, and the coordinates and brightness observation value of the corresponding point of the target point in at least one non-reference image; the brightness residual construction module 940 is used to... For each group of corresponding points, the imaging distance difference from the target point to each corresponding point is calculated. The observed brightness value of the target point is used as the reference brightness value. Based on the underwater imaging physical model, the initial water optical parameters, and each imaging distance difference, the theoretical brightness value of each corresponding point in the group is calculated. For each corresponding point, the brightness residual is calculated based on the theoretical brightness value and the observed brightness value of the corresponding point in the non-reference image. The parameter inversion module 950 is used to establish a target optimization function by minimizing the sum of squares of the brightness residuals of all corresponding points, and to iteratively solve the target optimization function to obtain the target water optical parameters. The target optimization function uses the red channel absorption characteristics as a regularization term to constrain the solution space range. The target water optical parameters include at least one of the following: water absorption coefficient, water scattering coefficient, water backscattering coefficient, background light intensity, and forward scattering blurring parameter.

[0085] In some embodiments, the parameter inversion module 950 is further configured to: introduce a physical regularization term during the iterative inversion process, using the non-negativity and value range of the red light band as constraints based on its strong absorption and weak scattering characteristics, while simultaneously constraining the range of blue-green channel parameters by utilizing the wavelength dependence of water absorption and scattering intensity; and incorporate the gradient of the regularization term into the parameter update during each iteration of optimization, so that the parameters converge toward the optimal solution fitted to the data while being guided by the physical constraints of the red channel absorption characteristics.

[0086] In some embodiments, the in-situ measurement system for water optical parameters further includes: a radiation uniformity module 960, used to acquire standard optical test card images at different underwater distances using the measuring device; to fit the radiation deviation between the non-reference imaging unit and the reference imaging unit one by one using the standard optical test card images at multiple distances to obtain radiation response correction coefficients, brightness compensation factors, and color correction matrices; and to perform radiation uniformity calibration on the plurality of imaging units using the radiation response correction coefficients, the brightness compensation factors, and the color correction matrix.

[0087] In some embodiments, the in-situ measurement system for optical parameters of water bodies further includes an imaging distance difference calculation module 970, which is used to calculate the imaging distance difference between different imaging units corresponding to the same target point based on the installation parameters, calibration results, corresponding point positions or auxiliary ranging information of the imaging unit.

[0088] In some embodiments, the in-situ measurement system for water body optical parameters further includes: a result output and application module 980, used to output the inverted optical parameters of the target water body, and for each image, input the optical parameters of the target water body into the underwater imaging physical model to obtain the imaging distance output by the underwater imaging physical model, and perform the following steps based on the optical parameters of the target water body and the imaging distance to reconstruct the image: based on the optical parameters of the target water body and the imaging distance, subtract the backscattering component from the image to obtain the corresponding first processed image; convert the first processed image to the frequency domain, and eliminate the backscattering component in the frequency domain. After removing the forward scattering component, the image is inverted back to the spatial domain to obtain a pure direct attenuation component with the forward scattering component removed. The pure direct attenuation component is multiplied by an attenuation compensation factor and color restored to obtain the restored second processed image. Based on the channel features of the blue and green channels, the geometric structure information of the blue and green channels is used as a ground truth reference to repair the edge and texture structure information of the second processed image in the red channel, resulting in a third processed image. Based on the channel features of the red channel, the red channel is used as a scattering ground truth reference to correct the scattering noise in the blue and green channels of the third processed image, resulting in a fourth processed image.

[0089] In some embodiments, the optical parameters of the target water body include at least one of the following: water absorption coefficient; water scattering coefficient; water backscattering coefficient; background light intensity; forward scattering blurring parameter; optical attenuation parameter under different bands or different color channels; and equivalent optical parameters related to water turbidity, suspended particulate matter concentration, or water color.

[0090] The in-situ measurement system for optical parameters of water provided in this application can realize the various processes implemented in the method embodiments shown above for in-situ measurement of optical parameters of water. To avoid repetition, these processes will not be described again here.

[0091] The in-situ measurement system for optical parameters of water bodies in this application embodiment can be a device, or a component, integrated circuit, or chip in an electronic device. This application embodiment does not impose specific limitations.

[0092] The in-situ measurement system for optical parameters of water bodies in this embodiment can be a device with an operating system. This operating system can be Android, iOS, or other possible operating systems; this embodiment does not specifically limit its application.

[0093] like Figure 11 As shown, this application embodiment also provides an electronic device 1100, including a processor 1110 and a memory 1120. The memory 1120 stores a program or instructions that can run on the processor 1110. When the program or instructions are executed by the processor 1110, they implement the various processes of the above-described embodiment of the in-situ measurement method for optical parameters of water bodies and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0094] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described in-situ measurement method for optical parameters of water bodies and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0095] The processor mentioned above is the processor in the terminal described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk. In some examples, the readable storage medium may be a non-transient readable storage medium.

[0096] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the various processes of the above-described embodiment of the in-situ measurement method for optical parameters of water bodies, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0097] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0098] This application also provides a computer program / program product, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions, which, when executed by a computer, cause the computer to perform the various processes shown in the above-described embodiment of the in-situ measurement method for optical parameters of water bodies, and achieve the same technical effect. To avoid repetition, these will not be described again here.

[0099] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on 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.

[0100] From the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of computer software products plus necessary general-purpose hardware platforms, and of course, they can also be implemented by hardware. The computer software product is stored in a storage medium (such as ROM, RAM, magnetic disk, optical disk, etc.) and includes several instructions to cause the terminal or network-side device to execute the methods described in the various embodiments of this application.

[0101] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other implementations under the guidance of this application without departing from the spirit and scope of the claims. All of these implementations are within the protection scope of this application.

Claims

1. A method for in-situ measurement of optical parameters of water bodies, characterized in that, include: Underwater images are acquired synchronously using a measurement device consisting of at least two non-collinear and non-coplanar imaging units whose relative poses are determined. The imaging distances from multiple imaging units to the same point in the underwater region are different. The plurality of imaging units are divided into a reference imaging unit and at least one non-reference imaging unit, and the image acquired by the reference imaging unit is used as the reference image and the image acquired by the non-reference imaging unit is used as the non-reference image. Obtain multiple sets of corresponding points between the reference image and at least one of the non-reference images, wherein each set of corresponding points includes the coordinates and brightness observations of a target point in the reference image, and the coordinates and brightness observations of the target point's corresponding points in at least one of the non-reference images; For each group of corresponding points, the imaging distance difference from the target point to each corresponding point is calculated. The observed brightness value of the target point is used as the reference brightness value. Based on the underwater imaging physical model, the initial water optical parameters, and each imaging distance difference, the theoretical brightness value of each corresponding point in the group is calculated. For each corresponding point, the brightness residual is calculated based on the theoretical brightness value and the observed brightness value of the corresponding point in the non-reference image. The underwater imaging physical model is implemented based on the Jaffe-McGlamery model, and the initial water optical parameters are selected using prior knowledge to guide the initial value selection. An objective optimization function is established to minimize the sum of squares of the brightness residuals of all the corresponding points, and the objective optimization function is iteratively solved to obtain the optical parameters of the target water body. The objective optimization function uses the red channel absorption characteristics as a regularization term to constrain the solution space range. The optical parameters of the target water body include at least one of the following: water body absorption coefficient, water body scattering coefficient, water body backscattering coefficient, background light intensity, and forward scattering blurring parameter. The objective optimization function uses the red channel absorption characteristics as a regularization term to constrain the solution space range. This includes: introducing a physical law regularization term during the iterative inversion process, using the non-negativity and value range of the red light band as constraints based on its strong absorption and weak scattering characteristics, while simultaneously using the wavelength dependence of water absorption and scattering intensity to constrain the blue and green channel parameter ranges; and incorporating the gradient of the regularization term into the parameter update during each iteration of optimization, so that the parameters converge toward the optimal solution fitted to the data while being guided by the physical constraints of the red channel absorption characteristics.

2. The method according to claim 1, characterized in that, Prior to synchronously acquiring underwater images using a measurement device composed of at least two non-collinear and non-coplanar imaging units determined by their relative poses, the method further includes: The measuring device is used to acquire images of standard optical test cards at different underwater distances; By fitting the radiation deviation between the non-reference imaging unit and the reference imaging unit one by one using the standard optical test card images at multiple distances, the radiation response correction coefficient, brightness compensation factor and color correction matrix are obtained; The radiometric consistency calibration of the plurality of imaging units is performed using the radiometric response correction coefficient, the luminance compensation factor, and the color correction matrix.

3. An in-situ measurement system for optical parameters of water bodies, characterized in that, include: An imaging unit module is a measuring device consisting of at least two non-collinear and non-coplanar imaging units whose relative poses are determined. The imaging distances of multiple imaging units to the same point in the underwater region are different. A synchronous acquisition control module is used to control the multiple imaging units to synchronously acquire underwater images; The corresponding point registration module is used to divide the plurality of imaging units into a reference imaging unit and at least one non-reference imaging unit, and to use the image acquired by the reference imaging unit as the reference image and the image acquired by the non-reference imaging unit as the non-reference image; to obtain multiple sets of corresponding points between the reference image and at least one of the non-reference images, wherein each set of corresponding points includes the coordinates and brightness observation values ​​of a target point in the reference image, and the coordinates and brightness observation values ​​of the corresponding points of the target point in at least one of the non-reference images; A brightness residual construction module is used to calculate the imaging distance difference from the target point to each of the corresponding points in each group, using the observed brightness value of the target point as a reference brightness value, and calculating the theoretical brightness value of each of the corresponding points in the group based on the underwater imaging physical model, initial water optical parameters, and each imaging distance difference; and for each corresponding point, calculating the brightness residual based on the theoretical brightness value and the observed brightness value of the corresponding point in the non-reference image, wherein the underwater imaging physical model is implemented based on the Jaffe-McGlamery model, and the initial water optical parameters are selected using prior knowledge to guide the initial value selection; The parameter inversion module is used to establish a target optimization function by minimizing the sum of squares of the brightness residuals of all the corresponding points, and to iteratively solve the target optimization function to obtain the target water body optical parameters. The target optimization function uses the red channel absorption characteristics as a regularization term to constrain the solution space range. The target water body optical parameters include at least one of the following: water body absorption coefficient, water body scattering coefficient, water body backscattering coefficient, background light intensity, and forward scattering blurring parameter. The parameter inversion module is also used to: introduce a physical regularization term during the iterative inversion process, using the non-negativity and value range of the red light band as constraints based on its strong absorption and weak scattering characteristics, while using the wavelength dependence of water absorption and scattering intensity to constrain the range of blue and green channel parameters; and incorporate the gradient of the regularization term into the parameter update during each iteration of optimization, so that the parameters are guided by the physical constraints of the red channel absorption characteristics while converging toward the optimal solution fitted to the data.

4. The in-situ measurement system for optical parameters of water bodies according to claim 3, characterized in that, Also includes: A radiation homogenization module is used to acquire images of standard optical test cards at different underwater distances using the measuring device; By fitting the radiation deviation between the non-reference imaging unit and the reference imaging unit one by one using the standard optical test card images at multiple distances, the radiation response correction coefficient, brightness compensation factor and color correction matrix are obtained; The radiometric consistency calibration of the plurality of imaging units is performed using the radiometric response correction coefficient, the luminance compensation factor, and the color correction matrix.

Images