A water resource information acquisition method based on remote sensing images

Abstract

The application discloses a water resource information acquisition method based on remote sensing images and belongs to the technical field of image recognition, and specifically comprises the following steps: firstly, remote sensing image data of a target water area is acquired, and a dynamic substrate spectrum template library is constructed by collecting measured spectrum reflection curves of different substrate types in a water-free state; then, for each spatial position in the image, the exposed substrate type of the adjacent area is matched according to the geographic coordinates, the corresponding spectrum features are extracted from the template library as the basic reflection reference; then, a light propagation path separation model is established to simulate the process that sunlight penetrates through the water layer, is reflected by the substrate and returns to the sensor, and the substrate reflection contribution component is generated in combination with the basic reflection reference; finally, the contribution component is deducted from the spectral reflectance to obtain the pure water body optical response value, and according to the physical correlation model between the optical response value and the water body transparency, a water body transparency parameter distribution map of each spatial position of the target water area is output; the application realizes reliable remote sensing acquisition of the water body transparency of a shallow water area.

Description

Technical Field

[0001] This invention relates to the field of image recognition technology, and more specifically to a method for collecting water resource information based on remote sensing images. Background Technology

[0002] Water resources are a core element for maintaining ecological security and human development. Obtaining key parameters such as water transparency is of great significance for water environment protection, water conservancy management, and ecological assessment. Remote sensing technology, with its advantage of large-scale synchronous observation, has become a core means of water resource information collection. Its inversion of water transparency relies on the accurate extraction of the inherent signals of the water body from spectral reflectance. However, in complex water areas such as shallow waters and nearshore zones, the spatial heterogeneity and dynamic changes of sediment types significantly interfere with spectral signals, limiting the accuracy of traditional remote sensing methods.

[0003] Existing remote sensing inversion methods are mostly based on statistical or physical models of spectral reflectance and transparency, estimating parameters through band combinations or optical equations. To address seabed interference, some techniques employ preset thresholds or homogenization assumptions to separate signals, such as dividing regions based on water depth and empirically correcting them. Other studies have established seabed spectral libraries to match typical characteristics; however, these methods generally ignore the heterogeneity of seabed distribution—such as the spatial mixing of sandy, muddy, and gravelly materials—and temporal changes such as those caused by water erosion and tides. They also fail to fully consider the complex effects of light coupling and propagation at the water-seabed interface on the spectrum.

[0004] In shallow water areas, sedimentary materials such as sandy and muddy bottoms are significantly affected by sedimentary environment and water erosion, resulting in substantial differences in particle composition, organic matter content, and surface condition. This leads to complex variations in the reflectance spectra of sediments in different areas, with sandy sediments exhibiting significantly higher reflectance than muddy sediments. These sediment reflectance signals, combined with water scattering and absorption signals, form a mixed spectrum. However, existing models assume uniform and stable optical properties of the sediments, failing to effectively characterize the heterogeneity within the same water body, such as the intermingling of different sediment types and variations in particle size gradients. This results in the inability to accurately separate sediment reflectance from the inherent signals of the water body during transparency inversion. For example, adjacent high-reflectance sandy and low-reflectance muddy sediments may be misinterpreted by traditional models as differences in water turbidity, leading to systematic biases in the inversion results and affecting the reliability of water quality assessment and ecological monitoring in shallow water areas. Summary of the Invention

[0005] The purpose of this invention is to provide a method for collecting water resource information based on remote sensing imagery, and to solve the following technical problems:

[0006] Existing inversion models typically assume that the optical properties of the substrate are uniform and stable, neglecting the heterogeneity of the substrate in spatial distribution, which leads to deviations in the inversion results of water transparency.

[0007] The objective of this invention can be achieved through the following technical solutions:

[0008] A method for collecting water resources information based on remote sensing imagery includes the following steps:

[0009] Acquire remote sensing image data of the target water area, wherein the remote sensing image data includes spectral reflectance corresponding to multiple spatial locations;

[0010] A dynamic sediment spectral template library is constructed. The dynamic sediment spectral template library is established by collecting measured spectral reflectance curves of different sediment types under anhydrous conditions. Each sediment type corresponds to an independent spectral feature dataset.

[0011] For each spatial location in the remote sensing image, the exposed substrate type of the adjacent area is matched according to the geographic coordinates of the spatial location, and the spectral feature dataset of the corresponding substrate type is extracted from the dynamic substrate spectral template library as the basic reflectance reference.

[0012] A light propagation path separation model is established, which simulates the complete path of sunlight penetrating the water layer, reaching the bottom surface, and returning to the sensor after secondary reflection. The bottom reflection contribution component is generated by combining the basic reflection reference.

[0013] The optical response value of pure water is obtained by subtracting the bottom sediment reflection contribution component from the spectral reflectance of remote sensing image data. Based on the physical correlation model between the optical response value of pure water and water transparency, a water transparency parameter distribution map of the spatial location of the target water area is output.

[0014] As a further aspect of the present invention: the process of constructing the dynamic substrate spectral template library is as follows:

[0015] During the dry season of the target water area, field surveys were conducted, and shallow areas with water levels below the natural threshold were selected as sampling areas for exposed sediment. The exposed sediment sampling areas were divided into grids, and physical samples of surface sediments were collected from each grid unit and the geographic coordinates of the physical samples were recorded.

[0016] The continuous spectral reflectance curves of physical samples in the visible to near-infrared bands were measured under dry conditions. Based on the sediment particle size distribution and organic matter content, the physical samples were classified into sandy substrate, muddy substrate, gravelly substrate, and mixed substrate types. Cluster analysis was performed on the spectral reflectance curves of substrate samples of the same type, and abnormal spectral reflectance curves with dispersion exceeding the set threshold were removed.

[0017] Calculate the band-weighted average of the retained spectral reflectance curves to form a standardized spectral reflectance set for the substrate type in a set band; associate and store the standardized spectral reflectance set with the corresponding geographic coordinates to form a spatially indexed dynamic substrate spectral template library.

[0018] As a further aspect of the present invention: the process of matching the exposed substrate type of adjacent regions is as follows:

[0019] A search buffer is established centered on the current location to be processed, and the radius of the search buffer is determined based on the historical maximum receding range of the water area; all exposed bottom sampling points located within the search buffer are retrieved from the dynamic bottom spectral template library.

[0020] Calculate the surface distance and elevation difference between the exposed substrate sampling point and the current spatial location, and exclude exposed substrate sampling points whose elevation difference exceeds the terrain undulation threshold; count the frequency of occurrence of substrate types for the remaining exposed substrate sampling points, and select the substrate type with the highest frequency as the matching result; when multiple substrate types have the same frequency, select the substrate type of the exposed substrate sampling point with the smallest surface distance from the current spatial location.

[0021] If there are no exposed substrate sampling points in the search buffer, the search is expanded to the typical substrate distribution map of the watershed to which the target water area belongs for analog matching.

[0022] As a further aspect of the present invention: the process by which the light propagation path separation model generates the substrate reflection contribution component is as follows:

[0023] Based on the imaging time of remote sensing images and satellite orbit, the incident angle of sunlight and the observation angle of the sensor are calculated. The propagation process of incident light penetrating the water layer and reaching the bottom surface is simulated, and the first propagation attenuation is calculated. The propagation process of light reflected from the bottom layer penetrating the water layer and reaching the sensor is simulated, and the second propagation attenuation is calculated.

[0024] The first and second propagation attenuation values ​​are applied to the basic reflection reference value to generate the bottom sediment reflection contribution component; a water surface specular reflection correction factor is introduced to dynamically adjust the bottom sediment reflection contribution component. The water surface specular reflection correction factor is generated by combining wind and wave disturbance model with real-time meteorological data.

[0025] As a further aspect of the present invention: the calculation process for propagation attenuation is as follows:

[0026] A region in the target water area with a water depth greater than the optical depth was selected as a pure water reference area; the spectral reflectance of the pure water reference area in the remote sensing image was extracted as the pure water signal without bottom sediment interference.

[0027] Establish the parameter fitting relationship between pure water body signals and theoretical water body absorption models, wherein the theoretical water body absorption models include phytoplankton pigment absorption bands, suspended particulate matter absorption bands, and dissolved organic matter absorption bands;

[0028] The parameters of the theoretical water absorption model are adjusted by nonlinear optimization algorithm to minimize the root mean square error between the simulated and measured spectral curves. The optimized parameters of the theoretical water absorption model are then used as the benchmark values ​​for the spectral absorption characteristics of the current water body.

[0029] As a further aspect of the present invention: the specific selection process for the pure water reference area is as follows:

[0030] The stability of the spectral reflectance of the selected pure water reference area is analyzed in the time series. When the deviation between the reflectance change trend of a specific band in the reference area and the inversion result of the physical correlation model of water transparency in multiple consecutive remote sensing images continues to increase, the reference area reselection mechanism is activated.

[0031] Based on the spatial distribution pattern of historical propagation attenuation, the optical stability index of each deep water area in the target water area is calculated. The optical stability index reflects the correlation between water depth and bottom sediment reflection suppression ability. The candidate area with the highest optical stability index is selected as the reference area for the new pure water body, ensuring that the water depth of the candidate area is greater than that of the original reference area and the fluctuation range of historical propagation attenuation is lower than the set tolerance threshold.

[0032] Extract the spectral reflectance of the new pure water reference area, refit the parameters of the theoretical water absorption model, and update the optimized theoretical water absorption model parameters to the spectral absorption characteristic baseline value of the light propagation path separation model.

[0033] As a further aspect of the present invention: the calculation process for the optical response value of the pure water body is as follows:

[0034] Obtain the total spectral reflectance of remote sensing images in a set band, calculate the bottom sediment reflection contribution component at the current spatial location, and determine the direct reflection component of the water surface.

[0035] The optical response value of pure water is obtained by subtracting the contribution component of sediment reflection and the direct reflection component of water surface from the total spectral reflectance value.

[0036] When the optical response value of pure water is lower than the preset noise threshold, spatial interpolation compensation is performed using the attenuation gradient of the nearby spatial location, and atmospheric scattering effect correction is applied to the optical response value of pure water after spatial interpolation compensation.

[0037] As a further aspect of the present invention: the construction process of the physical association model is as follows:

[0038] A fixed monitoring buoy array was deployed in the target water area to simultaneously acquire on-site measurements of water transparency at different spatial locations, and the pure water optical response values ​​at the corresponding spatial locations at the time of remote sensing image acquisition were extracted.

[0039] The statistical correlation between the specific blue-green band ratio of the optical response value of pure water and the field measurement value of water transparency was analyzed; a piecewise linear regression equation was constructed, in which the first slope linear relationship was used when the optical response value of pure water was in the low value range, and the second slope linear relationship was used when the optical response value of pure water was in the high value range.

[0040] The calibration sub-regions were divided according to hydrogeological units, and the parameters of the piecewise linear regression equation were optimized separately for each calibration sub-region.

[0041] As a further aspect of the present invention, it also includes acquiring multiple periods of remote sensing image data of the target water area in a continuous time series, independently generating a water transparency parameter distribution map for each period of remote sensing image data, and establishing a spatiotemporal change analysis matrix to detect abrupt changes in water transparency at the same spatial location.

[0042] When the duration of a mutation point exceeds a set period and the magnitude of the change exceeds the natural fluctuation threshold, the bottom sediment spectral template library update process is triggered.

[0043] The process of updating the substrate spectral template library includes re-collecting exposed substrate samples at spatial locations, measuring the spectral reflectance curves of the new exposed substrate samples and updating the dynamic substrate spectral template library data, recalculating the water transparency parameters for historical periods using the updated dynamic substrate spectral template library, and generating a time series dataset after consistency correction.

[0044] The beneficial effects of this invention are:

[0045] This invention constructs a dynamic sediment spectral template library and combines it with the matching of exposed sediment types in adjacent areas to accurately obtain the spectral characteristic benchmark of the sediment, thus solving the problem of inaccurate reflection reference caused by the spatiotemporal heterogeneity of sediment types. By simulating the complete path of light penetration through the water layer, sediment reflection, and return to the sensor through a light propagation path separation model, it accurately generates the sediment reflection contribution component, effectively separating sediment reflection from the inherent optical signal of the water body, overcoming the defect of existing models that assume uniform sediment and cannot quantify sediment contribution. By subtracting the sediment reflection contribution component from the spectral reflectance to obtain the optical response value of the pure water body, and combining it with the physical correlation model to invert transparency, it significantly reduces the inversion bias caused by sediment interference. At the same time, through time series analysis and a dynamic template library update mechanism, it can adapt to changes in sediment type with hydrological conditions, ensuring the spatial consistency and temporal stability of water transparency parameters in long-term monitoring, and ultimately achieving high-precision and high-reliability remote sensing acquisition of water transparency in shallow water areas. Attached Figure Description

[0046] The invention will now be further described with reference to the accompanying drawings.

[0047] Figure 1 This is a flowchart illustrating the present invention. Detailed Implementation

[0048] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] Please see Figure 1 As shown, this invention is a method for collecting water resource information based on remote sensing imagery, comprising the following steps:

[0050] First, remote sensing image data of the target water area is acquired. This type of data covers multiple bands from visible light to near-infrared, and records the spectral reflectance information of multiple spatial locations within the target water area at different wavelengths.

[0051] When constructing a dynamic sediment spectral template library, shallow, exposed sediment areas with water levels below the natural threshold during the dry season were selected as sampling areas. Surface sediment samples were collected in a grid pattern, and their geographic coordinates were recorded. Continuous spectral reflectance of the samples under dry conditions was measured. Based on grain size distribution, organic matter content, and other factors, the samples were classified into sandy, muddy, and other types. Abnormal spectra were removed through cluster analysis, and the band-weighted average of the retained spectra was calculated to form a standardized spectral set, which was then stored in association with geographic coordinates. This constructed a dynamic template library with spatial indexing to adapt to spatiotemporal changes in the sediment.

[0052] For each spatial location in the image, a search buffer is set with its coordinates as the center and based on the historical maximum receding range. Exposed sediment sampling points within the buffer are retrieved. After excluding outliers with excessive elevation differences, the frequency of sediment type among the remaining sampling points is counted, and the type with the highest frequency or the closest distance is selected as the matching result. If there is no data in the buffer, the spectral characteristics of the corresponding sediment type are obtained by referring to the typical sediment distribution map of the watershed as the basic reflectance reference.

[0053] A light propagation path separation model is established. Based on the image imaging time and satellite orbit parameters, the solar incidence and sensor observation angles are determined. The process of light penetrating the water layer to reach the bottom surface, being reflected, and then penetrating the water layer again to return to the sensor is simulated. The contribution of bottom reflection to the spectrum is quantified by combining the basic reflection reference, and the bottom reflection contribution component is generated.

[0054] The optical response value of pure water is obtained by subtracting the contribution of sediment reflection from the spectral reflectance. By setting up fixed monitoring points in the target water area, the measured values ​​of transparency and the corresponding optical response values ​​of pure water at the locations are acquired simultaneously. A physical correlation model between the two is established, and the optical response values ​​are converted into transparency parameters. Finally, a transparency parameter distribution map of each spatial location in the target water area is output.

[0055] In a preferred embodiment of the present invention, the process of constructing the dynamic substrate spectral template library is as follows:

[0056] Field surveys were conducted during the dry season in the target water area. At this time, the water level drops below the natural threshold, exposing the bottom sediment in shallow areas, providing ideal conditions for direct collection of bottom sediment samples not covered by water. Representative shallow areas were selected as sampling areas for exposed bottom sediment. These areas needed to cover the main bottom sediment types that might be present in the target water area. For example, river mouths tend to deposit muddy bottom sediment, shallow lake centers often have sandy bottom sediment, and river mouths may contain gravelly and mixed bottom sediment. The sampling area was divided into grids, with the grid density set according to the spatial variability of bottom sediment types. A sparser grid was used in areas with gentle terrain and relatively uniform bottom sediment composition, while a denser grid was used in areas with complex terrain and mixed bottom sediment types, ensuring that the collected samples could comprehensively reflect the spatial heterogeneity of the bottom sediment in the target water area.

[0057] When collecting physical samples of surface sediments for each grid unit, a unified sampling standard was followed: undisturbed sampling tools were used to obtain surface sediments at a depth of 0-10 cm to avoid contamination by deeper sediments affecting sample representativeness; the geographic coordinates of the samples were recorded simultaneously, and high-precision positioning equipment was used to ensure sub-meter accuracy, providing a foundation for accurate correlation between spectral data and spatial location. The acquired physical samples were air-dried naturally in a laboratory environment to remove moisture interference with spectral measurements. Subsequently, a continuous spectral reflectance scan of the dried samples was performed in the visible to near-infrared bands using a spectral measurement device, typically covering a wavelength range of 400-2500 nm, to capture the differences in reflectance characteristics of different sediments across the entire spectral range. For example, sandy sediments have higher reflectance in the visible light band, while muddy sediments have lower reflectance in the near-infrared band.

[0058] The sample classification process is based on the physicochemical properties of the sediments: First, particle size analysis is used to determine the proportions of sand, silt, and clay in the sediments. Combined with organic matter content detection, methods such as loss on ignition are used to classify the samples into sandy, muddy, gravelly, and mixed sediment types. Sediments with a sand content exceeding 50% are defined as sandy, those with a clay content exceeding 30% as muddy, those with a gravel content exceeding 40% as gravelly, and those with a balanced distribution of various particle components as mixed sediments. For the spectral reflectance curves of samples of the same sediment type, cluster analysis algorithms are used for data cleaning. By calculating the Euclidean distance or correlation coefficient between spectral curves, abnormal spectral curves with significantly higher dispersion than the population mean are identified and removed. These abnormal data may originate from contamination during the sampling process or temporary errors in the measuring equipment; removing them improves the reliability of the spectral template.

[0059] The retained valid spectral curves are statistically processed to calculate the weighted average value of each band. The weighting factor is set according to the band response characteristics of the remote sensing sensor, such as enhancing the sensitivity to commonly used bands in remote sensing images (blue, green, red, and near-infrared). This forms a standardized spectral reflectance set for each sediment type in the set bands. Finally, the standardized spectral data is associated with and stored with the corresponding geographic coordinates to construct a dynamic sediment spectral template library with spatial indexing. This template library allows users to quickly retrieve sediment spectral characteristics of nearby areas by geographic coordinates and can be dynamically updated by subsequent supplementary field sampling data, ensuring that the sediment spectral information remains synchronized with changes in sediment type caused by natural processes such as water erosion and deposition in the target water area.

[0060] In another preferred embodiment of the present invention, the process of matching the exposed substrate type of adjacent areas is as follows:

[0061] For each spatial location to be processed in the remote sensing image, a search buffer is established centered on the geographic coordinates of that location. The radius of the buffer is determined based on the historical maximum receding range of the target water area. This range is derived by analyzing water level monitoring data of the target water area over many years or by simulating water level changes under different hydrological conditions using a watershed hydrological model. This ensures that the buffer can cover areas of seabed that may be exposed during the dry season, avoiding the omission of seabed samples due to water level fluctuations. All exposed seabed sampling points falling within the search buffer are retrieved from a dynamic seabed spectral template library. These sampling points contain historically collected seabed types and their corresponding spectral characteristics.

[0062] To eliminate the interference of topographic factors on substrate type matching, it is necessary to calculate the surface distance and elevation difference between each exposed substrate sampling point and the current spatial location. The surface distance is calculated using planar Euclidean distance, and the elevation difference is determined by comparing the altitude difference between the two. A topographic relief threshold of 5 meters is set, and sampling points with elevation differences exceeding the threshold are removed. This is because significant topographic differences may lead to systematic changes in substrate type. For example, high-altitude areas are mostly distributed with gravelly substrates, while low-altitude depressions are prone to depositing muddy substrates. The two have significant differences in optical properties, and failure to remove them will lead to biased matching results. The frequency of substrate types at the remaining valid sampling points is statistically analyzed, and the substrate type with the highest frequency is prioritized as the matching result for the current spatial location. This strategy reflects the principle of the dominance of substrate type in spatial distribution, assuming that the dominant substrate type in the same area has a dominant influence on the substrate characteristics of the current location.

[0063] When multiple sediment types have the same frequency, a distance-first strategy is adopted: the sediment type of the sampling point with the shortest surface distance from the current spatial location is selected, as sediment types within close proximity are considered to have higher spatial correlation, reducing matching ambiguities caused by overlapping sediment types. For example, in areas where sandy and muddy sediments are interspersed, the type of the sampling point closest to the current location is more likely to reflect the actual sediment conditions at that location. If there are no exposed sediment sampling points within the search buffer zone, such as in areas where monitoring is being conducted for the first time or in special areas following extreme hydrological events, the matching range is expanded to a typical sediment distribution map of the target watershed. This map is constructed based on historical monitoring data from watershed geological surveys and expert knowledge, detailing the dominant sediment types and their distribution patterns in different sub-watersheds. Through analogical analysis, considering factors such as similar geological formation and consistency of water transport paths, the most likely sediment type is determined, ensuring that the matching process can still be effective even in the event of missing data.

[0064] Throughout the matching process, high-precision geographic coordinate positioning provided the foundation for spatial analysis, dynamic adjustment of the buffer zone adapted to changes in hydrological conditions, a topographic filtering mechanism eliminated interference from irrelevant data, and a multi-level matching strategy covered the needs for sediment type identification from local to watershed scales. These techniques worked together to form a sediment type identification method adapted to complex terrain and hydrological conditions. This method not only utilized the accuracy of field sampling data but also incorporated watershed-scale sediment distribution patterns, effectively solving the matching challenges caused by the spatial heterogeneity of sediment types in shallow water areas. This laid a solid foundation for the subsequent accurate separation of sediment reflection signals.

[0065] In another preferred embodiment of the present invention, the process by which the light propagation path separation model generates the substrate reflection contribution component is as follows:

[0066] When generating the contribution component of sediment reflection, the solar incidence angle and sensor observation angle are first determined based on the imaging time and satellite orbit parameters of the remote sensing image. These angle parameters are calculated using ephemeris data from the satellite platform and surface positioning information, and are used to construct a geometric model of light propagation. Next, the propagation process of incident light entering the water body from the atmosphere, penetrating the water layer, and reaching the sediment surface is simulated. During this process, the light is absorbed and scattered by components such as phytoplankton pigments, suspended particles, and dissolved organic matter in the water, resulting in energy attenuation; this attenuation is called the first propagation attenuation. Subsequently, the propagation process of light reflected from the sediment surface penetrating the water layer again and returning to the sensor is simulated. At this time, the light is also attenuated by the aforementioned components in the water, forming the second propagation attenuation. The simulation of both attenuation processes must consider the inherent optical characteristics of the water body, namely, the differences in the absorption and scattering coefficients of light of different wavelengths in water. These characteristics affect the degree of energy loss of light in the water layer.

[0067] The first and second propagation attenuation factors are applied to the previously obtained baseline reflection reference value, i.e., the standardized spectral reflectance corresponding to the substrate type. By calculating the energy loss of light in the water layer layer by layer, the contribution component of the substrate reflection signal to the sensor's received spectrum is generated. It is worth noting that the water surface condition has a significant impact on light propagation; therefore, a water surface specular reflection correction factor is introduced to dynamically adjust the substrate reflection contribution component. This correction factor is generated using a wind and wave disturbance model combined with real-time meteorological data. The wind and wave disturbance model simulates water surface roughness based on parameters such as wind speed and direction. The real-time meteorological data comes from ground meteorological stations or satellite inversion products and is used to assess changes in the intensity of water surface specular reflection, ensuring accurate correction of the interference of specular reflection on the substrate signal under different hydrodynamic conditions, such as calm water surface wave fluctuations.

[0068] In a preferred embodiment of this invention, the propagation attenuation is calculated as follows:

[0069] To accurately obtain the spectral absorption characteristics of water bodies, a region with a depth greater than the optical depth is first selected within the target water area as a pure water reference zone. The optical depth refers to the depth at which light energy decays to 1% of the surface layer; the seabed reflection signal outside this depth is negligible. Therefore, the selected reference zone ensures that it is not affected by seabed reflection. The spectral reflectance of the pure water reference zone is extracted from the remote sensing image as a pure water signal without seabed interference. This signal only contains the scattering and absorption characteristics of the water body itself and can truly reflect the inherent optical properties of the water body.

[0070] A parameter fitting relationship was established between the pure water body signal and the theoretical water body absorption model. The theoretical water body absorption model considers the absorption characteristics of components such as phytoplankton pigments, suspended particulate matter, and dissolved organic matter in different wavelength bands. These components correspond to specific absorption spectral ranges; for example, phytoplankton pigments mainly absorb the blue light band, suspended particulate matter has strong absorption in the green light band, and dissolved organic matter has significant absorption in the ultraviolet band. The model parameters were adjusted using a nonlinear optimization algorithm to minimize the root mean square error between the simulated spectral curve and the measured pure water body signal spectral curve. This process was achieved through iterative calculations. After each parameter adjustment, the difference between the simulated and measured values ​​was compared until the error converged to a preset accuracy range. Finally, the optimized model parameters were used as the benchmark values ​​for the spectral absorption characteristics of the current water body for subsequent propagation attenuation calculations, ensuring that the model accurately characterizes the optical absorption characteristics of the current water body.

[0071] In another preferred embodiment, the specific selection process for the pure water reference area is as follows:

[0072] To ensure the long-term validity of the spectral absorption characteristic benchmark values, a dynamic management mechanism for the pure water body reference area was designed. First, the stability of the spectral reflectance of the selected reference area was analyzed in a time series. Specifically, the reflectance variation trends of specific bands (such as blue and green light bands) in the reference area were compared with the inversion results of the water transparency physical correlation model across multiple consecutive remote sensing images. When the deviation between the two continuously increases, it indicates that the optical characteristics of the reference area may have changed, such as due to changes in water composition or the influence of exposed sediment. At this point, the reference area reselection mechanism is activated.

[0073] The reselection process calculates the optical stability index of each deep-water area in the target water body based on the spatial distribution pattern of historical propagation attenuation. This index comprehensively considers the correlation between water depth and the ability to suppress bottom sediment reflection; the greater the water depth, the smaller the impact of bottom sediment reflection, and the higher the optical stability. Specifically, the calculation combines historical data to statistically analyze the fluctuation range of propagation attenuation in each area; the smaller the fluctuation, the higher the stability. Selecting the candidate area with the highest optical stability index as the new pure water body reference area requires meeting two key conditions: first, the water depth of the candidate area must be greater than the original reference area to ensure effective suppression of bottom sediment reflection signals; second, the fluctuation range of historical propagation attenuation must be lower than a set tolerance threshold to ensure the stability of the water body's optical properties and avoid calculation deviations caused by fluctuations in the regional optical environment.

[0074] After determining the new reference area, its spectral reflectance is extracted and the parameters of the theoretical water absorption model are refitted. The parameters are then calibrated again using a nonlinear optimization algorithm. The optimized model parameters are then updated into the light propagation path separation model as new baseline values ​​for spectral absorption characteristics. This dynamic update mechanism ensures that the model can adapt to changes in the optical properties of the target water body caused by seasonal variations and water flow disturbances, maintaining the accuracy of propagation attenuation calculations and thus guaranteeing the long-term reliable calculation of the contribution component of sediment reflection.

[0075] In another preferred embodiment of the present invention, the calculation process of the optical response value of the pure water body is as follows:

[0076] When calculating the optical response of a pure water body, the total spectral reflectance of a set band is first extracted from remote sensing imagery. This total value includes multiple components such as the water body's own optical signal, the contribution of sediment reflection, and direct reflection from the water surface. Based on the previously constructed light propagation path separation model, the sediment reflection contribution component at the current spatial location is calculated. Simultaneously, by analyzing the relationship between the solar altitude angle and water surface roughness, the direct reflection component from the water surface is determined. This signal originates from the specular reflection of sunlight on the water surface and is unrelated to the inherent optical properties of the water body; therefore, it needs to be separated separately.

[0077] By sequentially subtracting the bottom sediment reflection contribution and the direct water surface reflection component from the total spectral reflectance, a pure water optical response value reflecting only the scattering and absorption characteristics of the water body is obtained. When the pure water optical response value at some spatial locations is lower than a preset noise threshold, it may be due to sensor noise or extreme water conditions. In this case, spatial interpolation compensation is performed using the attenuation gradient of the neighboring region. Using surrounding valid data points as a reference, a reasonable value for the target location is calculated according to distance weight to ensure the continuity of spatial data. After compensation, atmospheric scattering effect correction is applied to the pure water optical response value to eliminate the scattering interference of atmospheric molecules and aerosols on light, ultimately obtaining a pure signal that can be directly used for transparency inversion.

[0078] In another preferred embodiment of the present invention, the construction process of the physical association model is as follows:

[0079] When constructing a physical correlation model between the optical response value of pure water and water transparency, a fixed monitoring buoy array is deployed in the target water area. The buoys are distributed to cover areas with different water depths, substrate types, and pollution gradients to ensure data representativeness. The buoys simultaneously record the field measurements of water transparency using standard transparency measurement tools, and the geographic coordinates of the measurement locations are accurately marked. At the same time, the optical response values ​​of pure water at these locations at the time of remote sensing image acquisition are extracted to form a matching dataset of optical signals and transparency.

[0080] This study analyzes the statistical correlation between the ratio of specific blue-green bands in the optical response values ​​of pure water and the measured transparency values ​​in the dataset. The blue band is sensitive to suspended particulate matter in the water, while the green band is significantly affected by chlorophyll. The ratio of the two can comprehensively reflect the clarity of the water. Based on the correlation characteristics, a piecewise linear regression equation is constructed. When the optical response value of pure water is in the low range, corresponding to high turbidity, a linear relationship with the first slope is used; when it is in the high range, corresponding to high transparency, a linear relationship with the second slope is used to accommodate the nonlinear correlation between the optical signal and the actual value in different transparency ranges.

[0081] The target water area was divided into multiple calibration sub-regions based on hydrogeological units, such as mountain river sections, plain lake areas, and river deltas. The water composition and sediment background differ in each sub-region. The slope and intercept parameters of the piecewise linear regression equation were optimized separately for each sub-region. The model inversion values ​​were continuously adjusted by comparing the deviations between the inversion values ​​and the field measurements to ensure that the inversion results for each sub-region accurately reflect the local water characteristics.

[0082] In another preferred embodiment of the invention, to achieve long-term dynamic monitoring, multiple periods of remote sensing image data of the target water area are acquired over a continuous time series. The aforementioned steps are performed independently on each period of image data to generate a water transparency parameter distribution map for the corresponding time period. Based on these distribution maps, a spatiotemporal variation analysis matrix is ​​established. By comparing the transparency values ​​of the same spatial location at different time periods, abrupt change points are detected. Abrupt change points refer to areas where transparency fluctuates abnormally within a short period of time, which may originate from changes in substrate type, sudden pollution, or hydrological events.

[0083] When the duration of a mutation point exceeds a set period and the magnitude of the change exceeds the natural fluctuation threshold, it indicates that it is not an accidental disturbance, but a systematic change in the characteristics of the sediment or water body. At this time, the sediment spectral template library update process is triggered. During the update, exposed sediment samples in the area where the mutation point is located are re-collected, with priority given to exposed areas during the dry season. The spectral reflectance curves of the new samples are measured, and the old data at the corresponding positions in the template library are replaced. If the overall sediment type in the area has changed, the sampling range is expanded to cover the newly emerging sediment type.

[0084] After the template library is updated, the water transparency parameters for historical periods are recalculated. Using the updated sediment spectrum as a benchmark, errors caused by biased sediment assumptions in past data are corrected, generating a consistent time-series dataset. This mechanism ensures the comparability of long-term monitoring data, avoids the invalidation of historical data due to sediment changes, and provides a reliable basis for analyzing water resource evolution trends.

[0085] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A water resource information acquisition method based on remote sensing images, characterized in that, The method comprises the following steps: acquiring remote sensing image data of a target water area, the remote sensing image data containing spectral reflectance corresponding to a plurality of spatial positions; constructing a dynamic substrate spectral template library, the dynamic substrate spectral template library being established by collecting measured spectral reflectance curves of different substrate types in a water-free state, each substrate type corresponding to an independent spectral feature data set; for each spatial position in the remote sensing image, matching the exposed substrate type of the adjacent region according to the geographic coordinates of the spatial position, and extracting the spectral feature data set of the corresponding substrate type from the dynamic substrate spectral template library as a basic reflection reference; establishing a light propagation path separation model, the light propagation path separation model simulating the complete path of sunlight penetrating through the water layer to reach the substrate surface and then being reflected back to the sensor twice, and generating a substrate reflection contribution component in combination with the basic reflection reference; subtracting the substrate reflection contribution component from the spectral reflectance of the remote sensing image data to obtain a pure water body optical response value; and outputting a water body transparency parameter distribution map of the spatial position of the target water area according to a physical correlation model of the pure water body optical response value and the water body transparency; the process of matching the exposed substrate type of the adjacent region comprises the following steps: establishing a search buffer area with the current spatial position to be processed as the center, the radius of the search buffer area being determined according to the historical maximum water recession range of the water area; and retrieving all exposed substrate sampling points located in the search buffer area in the dynamic substrate spectral template library; calculating the ground surface distance and the elevation difference of the exposed substrate sampling points and the current spatial position, and excluding the exposed substrate sampling points with an elevation difference exceeding a terrain fluctuation threshold; performing frequency statistics on the substrate types of the remaining exposed substrate sampling points, and selecting the substrate type with the highest frequency as the matching result; when there are multiple substrate types with the same frequency, selecting the substrate type to which the exposed substrate sampling point with the smallest ground surface distance belongs; if there is no exposed substrate sampling point in the search buffer area, expanding to a typical substrate distribution map of a basin to which the target water area belongs for analogy matching; the process of generating the substrate reflection contribution component by the light propagation path separation model comprises the following steps: calculating the sunlight incidence angle and the sensor observation angle according to the imaging time of the remote sensing image and the satellite orbit, simulating the propagation process of the incident light penetrating through the water layer to reach the substrate surface, and calculating a first propagation attenuation amount; simulating the propagation process of the light reflected by the substrate penetrating through the water layer to reach the sensor, and calculating a second propagation attenuation amount; applying the first propagation attenuation amount and the second propagation attenuation amount to the basic reflection reference value to generate the substrate reflection contribution component; introducing a water surface specular reflection correction factor to dynamically adjust the substrate reflection contribution component, the water surface specular reflection correction factor being generated by a wind wave disturbance model in combination with real-time weather data; the calculation process of the propagation attenuation amount comprises the following steps: selecting a region with a water depth greater than an optical action depth in the target water area as a pure water body reference area; and extracting the spectral reflectance of the pure water body reference area in the remote sensing image as a pure water body signal without substrate interference; establishing a parameter fitting relationship between the pure water body signal and a theoretical water body absorption model, the theoretical water body absorption model containing phytoplankton pigment absorption bands, suspended particulate matter absorption bands and dissolved organic matter absorption bands; and The theoretical water body absorption model parameters are adjusted by a nonlinear optimization algorithm to minimize the root mean square error between the simulated spectral curve and the measured spectral curve, and the optimized theoretical water body absorption model parameters are taken as the spectral absorption characteristic reference value of the current water area. 2.The water resource information collection method based on remote sensing images according to claim 1, characterized in that, The process of constructing the dynamic substrate spectral template library is as follows: In the dry season of the target water area, the shallow area with water level lower than the natural threshold is selected as the exposed substrate sampling area; the exposed substrate sampling area is divided into grids, and the physical sample of the surface sediment of each grid unit is collected and the geographic coordinates of the physical sample are recorded; The visible light to near-infrared band continuous spectral reflectance curve of the physical sample in the dry state is measured; the physical sample is classified into sandy substrate type, muddy substrate type, gravel substrate type and mixed substrate type according to the sediment particle size distribution and organic matter content; the spectral reflectance curves of the same type of substrate sample are subjected to cluster analysis, and the abnormal spectral reflectance curves with a dispersion exceeding a set threshold are removed; The wave band weighted average value of the reserved spectral reflectance curve is calculated to form a standardized spectral reflectance set of the substrate type at a set wave band; the standardized spectral reflectance set and the corresponding geographic coordinates are stored in association to form a spatially indexed dynamic substrate spectral template library. 3.The water resource information collection method based on remote sensing images according to claim 1, characterized in that, The specific selection process of the pure water body reference area is as follows: The spectral reflectance stability of the selected pure water body reference area is analyzed in the time sequence; when the deviation of the specific wave band reflectance change trend of the reference area from the inversion result of the water body transparency physical correlation model continuously increases, the reference area reselection mechanism is activated; Based on the spatial distribution pattern of the historical propagation attenuation, the optical stability index of each deep water area of the target water area is calculated, which reflects the correlation between water depth and substrate reflection suppression ability; the candidate area with the highest optical stability index is selected as the new pure water body reference area, which ensures that the water depth of the candidate area is greater than that of the original reference area and the fluctuation amplitude of the historical propagation attenuation is lower than a set tolerance threshold; The spectral reflectance of the new pure water body reference area is extracted, the theoretical water body absorption model parameters are re-fitted, and the optimized theoretical water body absorption model parameters are updated to the spectral absorption characteristic reference value of the light propagation path separation model. 4.The water resource information collection method based on remote sensing images according to claim 1, characterized in that, The calculation process of the pure water body optical response value is as follows: The spectral reflectance total value of the remote sensing image at a set wave band is obtained, the substrate reflection contribution component at the current spatial position is calculated, and the water surface direct reflection component is determined; The pure water body optical response value is obtained by deducting the substrate reflection contribution component and the water surface direct reflection component from the spectral reflectance total value; When the pure water body optical response value is lower than a preset noise threshold, spatial interpolation compensation is performed using the attenuation gradient of the adjacent spatial position, and the spatially interpolated pure water body optical response value is subjected to atmospheric scattering effect correction. 5.The water resource information collection method based on remote sensing images according to claim 1, wherein, The construction process of the physical correlation model is as follows: A fixed monitoring buoy array is arranged in the target water area, and the water body transparency field measurement values at different spatial positions are synchronously obtained, and the pure water body optical response values at the remote sensing image acquisition time of the corresponding spatial positions are extracted; The statistical correlation between the specific blue-green band ratio of the optical response value of the pure water body and the measured value of the water transparency is analyzed, and a piecewise linear regression equation is constructed, which adopts a first slope linear relationship when the optical response value of the pure water body is in a low value interval, and adopts a second slope linear relationship when the optical response value of the pure water body is in a high value interval. The calibration sub-regions are divided according to hydrogeological units, and the parameters of the piecewise linear regression equation are optimized for each calibration sub-region. 6.The water resource information collection method based on remote sensing images according to claim 1, wherein, It also includes obtaining multi-period remote sensing image data of the target water area in a continuous time sequence, independently generating a water transparency parameter distribution map for each remote sensing image data, and establishing a space-time change analysis matrix to detect the mutation point of the water transparency at the same spatial position. When the duration of the mutation point exceeds the set period and the change amplitude exceeds the natural fluctuation threshold, the substrate spectrum template library updating process is triggered. The substrate spectrum template library updating process includes re-collecting the exposed substrate sample at the spatial position, measuring the spectral reflectance curve of the new exposed substrate sample and updating the dynamic substrate spectrum template library data, using the updated dynamic substrate spectrum template library to re-calculate the water transparency parameters of the historical period, and generating a consistent corrected time series data set.

Images