Water depth inversion method

Abstract

The invention belongs to the technical field of ocean water color remote sensing, and discloses a water depth inversion method, which comprises the following steps: acquiring satellite remote sensing data, tidal water level data and a sea chart of a research area; obtaining water-leaving radiance based on the satellite remote sensing data; carrying out vectorization processing on the sea chart to obtain water depth point data; matching the satellite remote sensing image data with the water depth point data, and extracting water leaving radiance and water depth information of a corresponding wave band; carrying out tide correction based on the imaging time of the satellite remote sensing image data in combination with the tide water level data to obtain corrected water depth information; constructing a water depth inversion model based on the off-water radiance and the corrected water depth information; and verifying the water depth inversion model, and performing water depth inversion of the research area based on the water depth inversion model passing the verification. According to the technical scheme, the advantages of various data can be fully utilized, the precision and reliability of water depth inversion are improved, and the method can be applied to all time periods.

Description

Technical Field

[0001] The invention belongs to the technical field of ocean color remote sensing, and in particular relates to a water depth inversion method. Background Art

[0002] Shallow sea depth data is an important basic geographic spatial information required for economic and military applications such as shallow sea environmental governance, shallow sea resource development and utilization, shallow sea navigation, and island operations. The change of the bottom boundary of a bay is reflected in the water depth. The change of historical water depth can directly reflect the land reclamation in the area. The change of bay water depth will also affect the change of the waterway. The change of water depth affects the tidal volume and water exchange capacity of the bay. The accuracy of the boundary field of general ocean numerical models is low because it is difficult to obtain the real water depth at the time of simulation. Similarly, when evaluating the change of a bay's disaster-bearing capacity, an accurate bottom boundary is necessary.

[0003] At present, the water depth is measured by acoustic means, which uses the emission, reception and speed of sound waves to obtain water depth data. Although this method can quickly obtain water depth data, it is time-consuming and labor-intensive, has a relatively low coverage, cannot update and obtain previous water depth data in a timely manner, and requires ships to go to the measurement area, which is affected by natural factors such as weather. Satellite remote sensing measurement provides a new way to invert water depth, which is mainly divided into active remote sensing and passive remote sensing. Airborne depth measurement LiDAR (Light Detection And Ranging) can take into account both accuracy and efficiency to achieve water depth detection around islands and reefs, but the data acquisition cost is high and is often restricted by airspace control.

[0004] At the same time, the traditional empirical formula model cannot invert the historical depth. The remote sensing reflection values ​​of the same satellite image passing through the same area are different, which makes the traditional empirical depth inversion model unsuitable for different time periods. This is also the reason why the research on general depth inversion stays at the steps of model application and model accuracy verification. In addition, historical depth is also difficult to obtain data. Except for some bathymetric data that will be preserved in the form of archives, most areas lack previous depth data. Satellite-borne Lidar bathymetric data can obtain years after 2018, but the scanning range of satellite-borne Lidar is strip-shaped, and the satellite revisit cycle is long. However, the development and transformation of a bay will inevitably bring changes to boundary conditions. Optical satellite remote sensing images can record the historical form of the bay, and the changes in the seabed benthic can be obtained through inversion. At present, CEBCO and ETOPO provide historical charts and water depth archives, but the resolution is too coarse, and medium-resolution data downloads are also provided. These data are more consistent with the actual water depth in Class I water bodies, but in Class II water bodies, there is a large error with the actual water depth results, especially the water depth of coastal waters and harbors, which makes it impossible to use them directly as the actual water depth. Summary of the invention

[0005] The purpose of the present invention is to provide a water depth inversion method to solve the problems existing in the above-mentioned prior art.

[0006] To achieve the above object, the present invention provides a water depth inversion method, comprising:

[0007] Obtain satellite remote sensing data, tidal level data and nautical charts for the study area;

[0008] Acquire water-leaving radiance based on the satellite remote sensing data;

[0009] Performing vector processing on the nautical chart to obtain water depth point data;

[0010] Matching the satellite remote sensing image data with the water depth point data to extract the water-leavage radiance and water depth information of the corresponding band;

[0011] Performing tidal correction based on the imaging time of the satellite remote sensing image data and the tidal water level data to obtain corrected water depth information;

[0012] Constructing a water depth inversion model based on the water-leaving radiance and the corrected water depth information;

[0013] The water depth inversion model is verified, and the correlation analysis is performed based on the tidal water level data and the water depth at different inversion times; the feasibility of the water depth algorithm is judged based on the trend and determination coefficient of the tidal water level and the inverted water depth. The water depth inversion of the study area is performed based on the verified water depth inversion model.

[0014] Optionally, obtaining water-leaving radiance based on the satellite remote sensing data specifically includes:

[0015] The satellite remote sensing data is subjected to radiation calibration and atmospheric correction to obtain the water-leaving radiance.

[0016] Optionally, before matching the satellite remote sensing image data with the water depth point data, the process further includes extracting water bodies from the satellite remote sensing data, and matching the satellite remote sensing data after the water body extraction with the water depth point data.

[0017] Optionally, water body extraction is performed on the satellite remote sensing data, and the specific calculation formula is:

[0018]

[0019] Where, NDWI represents the normalized difference water index; L Green With L NIR They are the water-leaving radiances of the green band and near-infrared band in satellite remote sensing images, respectively.

[0020] Optionally, tidal correction is performed based on the imaging time of the satellite remote sensing image data in combination with the tidal water level data, and the specific calculation formula is:

[0021] Z=Z chart +Z tide

[0022] Where Z is the corrected water depth information, Z chart is the chart water depth data, Z tide Tidal data.

[0023] Optionally, constructing a water depth inversion model based on the water-leaving radiance and the corrected water depth information specifically includes:

[0024] A random forest model / ratio model is constructed based on the water-leaving radiance and the corrected water depth information as a water depth inversion model.

[0025] Optionally, the training process of the random forest model includes:

[0026] An initial random forest model is constructed, and the water-leaving radiance is input as feature input data into the initial random forest model for water depth inversion. Training is performed with the goal of minimizing the loss between the initial training result after water depth inversion and the corrected water depth information corresponding to the water-leaving radiance to obtain a trained random forest model, and water depth inversion of the study area is performed based on the trained random forest model.

[0027] Optionally, the ratio model is specifically:

[0028] Z=A1*(L i / L j )+B1

[0029] Where Z represents the water depth, L i and L j represent the water-leaving radiances of the two selected bands respectively, and A1 and B1 are the coefficients of the model.

[0030] Optionally, verifying the water depth inversion model specifically includes:

[0031] Based on the water depth inversion model, the water depth inversion of the study area is carried out, and the water depth inversion data and the tide level data at the corresponding time are recorded;

[0032] Calculating the determination coefficient between the water depth inversion data and the tide level data at the corresponding time as the first verification result;

[0033] When the first verification result is less than a preset verification threshold, the water depth inversion model fails the verification;

[0034] When the first verification result is greater than a preset verification threshold, the water depth inversion model passes the verification. After passing the verification, the historical GEBCO data of the study area is obtained, and a correlation analysis is performed on the water depth inversion data and the historical GEBCO data, and the correlation analysis result is used as the second verification result;

[0035] When the second verification result is less than a preset verification threshold, the water depth inversion model fails the verification;

[0036] When the second verification result is greater than a preset verification threshold, the water depth inversion model passes the verification, and the water depth inversion of the study area is performed based on the verified water depth inversion model.

[0037] The technical effects of the present invention are:

[0038] The present invention is based on the phenomenon that the water level changes due to the rise and fall of seawater tides, the instantaneous water depth changes with the rise and fall of tides, the real water depth is not affected, and the instantaneous water depth time series of the inversion algorithm conforms to the corresponding water level time series change. After the remote sensing satellite image is used for radiation calibration and atmospheric correction, the water radiance is obtained. The water depth provided by the chart data is combined with the water depth of multiple phases and the water level data of the corresponding phases to fit the determination coefficient of the water level and water depth. The feasibility of the algorithm application at different times is verified by the trend and determination coefficient of the two. The multi-source data fusion method of the present invention can make full use of the advantages of various data, improve the accuracy and reliability of water depth inversion, and can verify the feasibility of the algorithm application in various time periods. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0040] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0041] Figure 1 It is a schematic diagram of the overall process in an embodiment of the present invention;

[0042] Figure 2 It is a sampling point in the Yangjiang sea area, which is one of the study areas in the embodiment of the present invention;

[0043] Figure 3 The sampling point of Zhanjiang Bay is one of the study areas in the embodiment of the present invention;

[0044] Figure 4 It is a time series diagram of the inverted water depth and water level at the sampling points in the Yangjiang sea area in the embodiment of the present invention;

[0045] Figure 5 It is a time series diagram of the inverted water depth and water level at the sampling point of Zhanjiang Bay in the embodiment of the present invention;

[0046] Figure 6 It is a diagram of absolute coefficients of inverted water depth and water level in Yangjiang sea area and Zhanjiang Bay in an embodiment of the present invention;

[0047] Figure 7 It is a time series diagram of random forest inversion water depth results in Yangjiang sea area (black frame area) in an embodiment of the present invention;

[0048] Figure 8 This is a comparison chart between the inverted water depth results of Yangjiang sea area in 2022 and GEBCO2022 in an embodiment of the present invention. DETAILED DESCRIPTION

[0049] Various exemplary embodiments of the present invention will now be described in detail. This detailed description should not be considered as limiting the present invention, but should be understood as a more detailed description of certain aspects, features, and embodiments of the present invention.

[0050] It should be understood that the terms described in the present invention are only for describing special embodiments and are not intended to limit the present invention. In addition, for the numerical range in the present invention, it should be understood that each intermediate value between the upper and lower limits of the scope is also specifically disclosed. Each smaller range between the intermediate value in any stated value or stated range and any other stated value or intermediate value in the described range is also included in the present invention. The upper and lower limits of these smaller ranges can be independently included or excluded in the scope.

[0051] Unless otherwise indicated, all technical and scientific terms used herein have the same meanings as those of ordinary skill in the art generally understood by the present invention. Although the present invention describes only preferred methods, any method similar or equivalent to that described herein may also be used in the implementation or testing of the present invention. All documents mentioned in this specification are incorporated by reference to disclose and describe the methods related to the documents. In the event of a conflict with any incorporated document, the content of this specification shall prevail.

[0052] It will be apparent to those skilled in the art that various modifications and variations may be made to the specific embodiments of the present invention description without departing from the scope or spirit of the present invention. Other embodiments derived from the present invention description will be apparent to those skilled in the art. The present application description and examples are exemplary only.

[0053] The words “include,” “including,” “have,” “contain,” etc. used in this article are open-ended terms, meaning including but not limited to.

[0054] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0055] Embodiment 1

[0056] like Figure 1 - Figure 8 As shown, in this embodiment, a water depth inversion method is provided, including: obtaining satellite remote sensing data, tidal water level data and nautical charts of the study area; obtaining water-leaving radiance based on the satellite remote sensing data; vectorizing the nautical chart to obtain water depth point data; matching the satellite remote sensing image data with the water depth point data to extract the water-leaving radiance and water depth information of the corresponding band; performing tidal correction based on the imaging time of the satellite remote sensing image data combined with the tidal water level data to obtain the corrected water depth information; constructing a water depth inversion model based on the water-leaving radiance and the corrected water depth information; verifying the water depth inversion model, and performing correlation analysis based on the tidal water level data and the water depth at different inversion times; judging the feasibility of the water depth algorithm based on the trend and determination coefficient of the tidal water level and the inverted water depth. The water depth inversion of the study area is performed based on the verified water depth inversion model.

[0057] In this embodiment, based on the phenomenon that the water level changes due to the rise and fall of the sea tide, the instantaneous water depth changes with the rise and fall of the tide, and the real water depth is not affected, the instantaneous water depth time series of the inversion algorithm conforms to the corresponding water level time series change, and the water-leaving radiance is obtained after radiation calibration and atmospheric correction using remote sensing satellite images. The water depth provided by the nautical chart data is combined with the inverted water depths of multiple phases and the water level data of the corresponding phases to fit the inversion model of water level and water depth. The multi-source data fusion method of this embodiment can make full use of the advantages of various data, improve the accuracy and reliability of water depth inversion, and can be applied to various time periods.

[0058] The present embodiment belongs to the field of ocean color remote sensing technology, and specifically relates to a shallow sea water depth inversion verification method based on water level time series changes. According to the phenomenon that the water level changes due to the rise and fall of seawater tides, the instantaneous water depth changes with the tidal rise and fall, and the real water depth is not affected, the instantaneous water depth time series of the inversion algorithm conforms to the corresponding water level time series changes, and the water-leaving radiance is obtained after radiation calibration and atmospheric correction using remote sensing satellite images, and the water depth provided by the chart data is used to select the blue light (band2) and red light (band4) bands with higher correlation to construct a ratio model, and the inversion model of water level and water depth is fitted by combining the inverted water depths of multiple time phases and the water level data of the corresponding time phases. Subsequently, during the verification process, the water depth inversion of the random forest was performed by inputting the water-leaving radiance of the seven bands of Landsat, and 21 points were randomly selected for verification. It was found that the changes in the water depth time series could be well synchronized with the changes in the water level time series. The obtained fitting equation was applied to four time periods, and the water depth data of GEBCO in 2022 was selected for verification. The correlation coefficient reached 0.72. In this embodiment, the initial water depth inversion algorithm can be any algorithm, and the water level data can be data provided by the tide gauge or data from the water level meter. A reference for water level-based water depth time series verification is established, which provides a water depth inversion verification method for remote areas, areas that cannot be mapped, and areas lacking measured water depth time series data.

[0059] like Figure 1 As shown, this embodiment provides a shallow sea water depth inversion verification method based on the change of water level time series. Based on the water-leav-the-water radiance of each band and the chart water depth data, a random forest model and a ratio model are established. The above models are applied to satellite images at different times to invert the instantaneous water depth at the time of satellite transit, and the water depth time series and tide level time series are obtained. The correlation coefficient of the two time series is calculated to verify the feasibility of the above algorithm.

[0060] Take the waters near Hailingshan Island in Yangjiang City, Guangdong Province and Zhanjiang Bay in Zhanjiang City as examples (e.g. Figure 2 and Figure 3 As shown in the figure, the original Landsat8 and Landsat9 hyperspectral satellite remote sensing images from 2015 to 2023 are obtained, and the cloud coverage rate is less than 30%. The water level data of the tidal station is the data at the time of satellite transit, namely Hailingshan Island Station and Zhanjiang Station. The names of the nautical charts are "Near Hailing Island" and "Donghai Island to Tiaoshun Island", and the chart numbers are 86101 and 88103.

[0061] The ENVI remote sensing image processing platform is used to preprocess the acquired Landsat raw satellite remote sensing data, including radiation calibration and atmospheric correction.

[0062] (1) Radiometric calibration: The signals received by satellite sensors are affected by the absorption, scattering, reflection, refraction and transmission of aerosols, cloud particles and water vapor in the atmosphere, which may cause distortion of remote sensing image information. Radiometric calibration is required to convert the image DN value into the apparent radiance value L.

[0063] L=Gain*DN+offset

[0064] Gain is the gain coefficient, Offset is the offset, and the general radiation calibration tool (RadiometricCalibration) in ENVI can automatically read parameters from the metadata file to complete the radiation calibration.

[0065] (2) Atmospheric correction: Atmospheric correction is mainly to eliminate the influence of factors such as atmosphere and light on the reflection of objects and obtain the real reflectivity data of objects. FLAASH atmospheric correction is used, and its principle is based on the atmospheric correction model of the non-uniform surface and Lambertian surface in the solar spectrum range:

[0066]

[0067] Where: L is the total radiation received by the satellite, ρ is the reflectivity of the pixel, and ρ e is the average reflectivity of the surrounding area; s is the reflectivity of the large balloon surface; L a is the amount of radiation backscattered by the atmosphere: A and B are coefficients that depend on the atmosphere and geometry.

[0068] The preprocessed Landsat satellite remote sensing data is clipped out of the water body by normalizing the water body index.

[0069] In order to prevent land pixels from participating in the water depth inversion calculation and improve the calculation efficiency, it is necessary to perform water-land segmentation, and the NDWI index is used for water-land segmentation. The calculation formula of the normalized difference water index is as follows:

[0070]

[0071] Where, NDWI represents the normalized difference water index; L Green With L NIR They represent the water-leaving radiance of the green band and the near-infrared band in the Landsat satellite remote sensing image respectively.

[0072] The water depth information provided by the nautical chart is used for vectorization to obtain water depth points after vectorization of the nautical chart; the steps include high-definition scanning, geo-referencing and water depth point vectorization.

[0073] (1) Use a scanner to scan the nautical chart to obtain a high-definition picture of the nautical chart.

[0074] (2) Use ArcGIS software for geo-referencing, select the latitude and longitude grid information in the scanned nautical chart, and align the chart by establishing control points and inputting the longitude and latitude provided on the nautical chart latitude and longitude grid.

[0075] (3) Create a sounding point feature and add a new field that contains lon (longitude), lat (latitude) and depth. Create a point based on the location of the sounding point provided by the nautical chart, then calculate the longitude and latitude and manually enter the depth.

[0076] Based on the clipped satellite remote sensing data and the water depth points after the nautical chart vectorization, the remote sensing image processing platform is used to extract the water radiance and water depth of each band corresponding to the water depth points after the nautical chart vectorization; and the data is matched with the satellite remote sensing image data set;

[0077] According to the transit time in the satellite remote sensing image source file, since the research area is located in China, the satellite transit time is standard universal time. Therefore, at the corresponding time 8 hours after the satellite transit time, the water depth point corresponding to the vectorized chart is calculated and tidal correction is performed to obtain the water depth data after tidal correction; and tidal correction is performed. The correction formula is as follows:

[0078] Z=Z chart +Z tide

[0079] Where Z represents the instantaneous water depth data; Z chart Indicates chart depth data; Z tide Represents tidal data. If the water depth data is less than 0m after tidal correction, it will be discarded.

[0080] This embodiment adopts two water depth inversion methods, namely, random forest model and ratio model.

[0081] The random forest model was applied to the Yangjiang waters, and the ratio model was applied to the Zhanjiang Bay.

[0082] (1) Establish a random forest model by inputting the water-leaving radiance of each band as the feature input and the water depth data after tidal correction, where the water-leaving radiance of each band is used as the feature and the water depth is used as the output result. The trained random forest model is applied to satellite images of different time periods to obtain the results of random forest model inversion of water depth. (2) Establish a ratio model: Z = A1*(L i / L j )+B1, by calculating the value of the band ratio and fitting it with the tide-corrected water depth data, A1 and B1 are obtained, and the fitted empirical formula is applied to satellite images of different time periods to obtain the results of the ratio model.

[0083] The sea area near the tidal station was randomly selected, and the remote sensing image processing platform was used to extract the water depth and tidal level data at the corresponding time at multiple points at different times (such as Figure 4 and Figure 5 ), calculate the coefficient of determination of two variables (as shown in Figure 6 As shown in Figure 2, the determination coefficient is used to determine whether the algorithm can be used to invert the water depth time series. According to the determination coefficient of the inverted water depth and water level, the R 2 The average is 0.71, and the R 2 The average is 0.44, and the Yangjiang sea area has a higher determination coefficient, so the random forest inversion algorithm in the Yangjiang sea area is suitable for inverting the water depth time series (such as Figure 7 ), using the 2022 water depth data provided by the GEBCO water depth archive, the Yangjiang waters (as shown Figure 2 The results of random forest model inversion (black box area shown) were correlated with GEBCO data. 2 Reach 0.77 (such as Figure 8 As shown), it is shown that the random forest model can be used for the inversion of water depth time series. It can be seen that this embodiment provides a verification of a water depth inversion method.

[0084] The above is only a preferred specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A water depth inversion method, characterized in that: include: Obtain satellite remote sensing data, tidal level data and nautical charts for the study area; Acquire water-leaving radiance based on the satellite remote sensing data; Performing vector processing on the nautical chart to obtain water depth point data; Matching the satellite remote sensing image data with the water depth point data to extract the water-leavage radiance and water depth information of the corresponding band; Performing tidal correction based on the imaging time of the satellite remote sensing image data and the tidal water level data to obtain corrected water depth information; Constructing a water depth inversion model based on the water-leaving radiance and the corrected water depth information; The water depth inversion model is verified, and the water depth inversion of the study area is performed based on the verified water depth inversion model.

2. A water depth inversion method according to claim 1, characterized in that: The method of obtaining water-leaving radiance based on the satellite remote sensing data specifically includes: The satellite remote sensing data is subjected to radiation calibration and atmospheric correction to obtain the water-leaving radiance.

3. A water depth inversion method according to claim 1, characterized in that: Before matching the satellite remote sensing image data with the water depth point data, the method further includes extracting water bodies from the satellite remote sensing data, and matching the satellite remote sensing data after the water body extraction with the water depth point data.

4. A water depth inversion method according to claim 3, characterized in that: Water body extraction is performed on the satellite remote sensing data, and the specific calculation formula is: Where, NDWI represents the normalized difference water index; L Green With L NIR They are the water-leaving radiances of the green band and near-infrared band in satellite remote sensing images, respectively.

5. A water depth inversion method according to claim 1, characterized in that: Tidal correction is performed based on the imaging time of the satellite remote sensing image data combined with the tidal water level data. The specific calculation formula is: Z=Z chart +Z tide Where Z is the corrected water depth information, Z chart is the chart water depth data, Z tide Tidal data.

6. A water depth inversion method according to claim 1, characterized in that: A water depth inversion model is constructed based on the water-leaving radiance and the corrected water depth information, specifically comprising: A random forest model / ratio model is constructed based on the water-leaving radiance and the corrected water depth information as a water depth inversion model.

7. A water depth inversion method according to claim 6, characterized in that: The training process of the random forest model includes: An initial random forest model is constructed, and the water-leaving radiance is input as feature input data into the initial random forest model for water depth inversion. Training is performed with the goal of minimizing the loss between the initial training result after water depth inversion and the corrected water depth information corresponding to the water-leaving radiance to obtain a trained random forest model, and water depth inversion of the study area is performed based on the trained random forest model.

8. A water depth inversion method according to claim 6, characterized in that: The ratio model is specifically: Z=A1*(L i / L j )+B1 Where Z represents the water depth, L i and L j represent the water-leaving radiances of the two selected bands respectively, and A1 and B1 are the coefficients of the model.

9. A water depth inversion method according to claim 1, characterized in that: The water depth inversion model is verified, specifically including: Based on the water depth inversion model, the water depth inversion of the study area is carried out, and the water depth inversion data and the tide level data at the corresponding time are recorded; Calculating the determination coefficient between the water depth inversion data and the tide level data at the corresponding time as the first verification result; When the first verification result is less than a preset verification threshold, the water depth inversion model fails the verification; When the first verification result is greater than a preset verification threshold, the water depth inversion model passes the verification. After passing the verification, the historical GEBCO data of the study area is obtained, and a correlation analysis is performed on the water depth inversion data and the historical GEBCO data, and the correlation analysis result is used as the second verification result; When the second verification result is less than a preset verification threshold, the water depth inversion model fails the verification; When the second verification result is greater than a preset verification threshold, the water depth inversion model passes the verification, and the water depth inversion of the study area is performed based on the verified water depth inversion model.

Images