CN115761463A - Shallow sea water depth inversion method, system, equipment and medium - Google Patents
Shallow sea water depth inversion method, system, equipment and medium Download PDFInfo
- Publication number
- CN115761463A CN115761463A CN202211362348.6A CN202211362348A CN115761463A CN 115761463 A CN115761463 A CN 115761463A CN 202211362348 A CN202211362348 A CN 202211362348A CN 115761463 A CN115761463 A CN 115761463A
- Authority
- CN
- China
- Prior art keywords
- water depth
- data
- remote sensing
- wave band
- selecting
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/30—Assessment of water resources
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a shallow sea water depth inversion method, which comprises the following steps: carrying out radiometric calibration, atmospheric correction and geometric correction on the original remote sensing image data, and carrying out water-land separation on the image. And inquiring a tide table, and preprocessing the water depth data. And taking logarithm of each wave band value of the extracted remote sensing image wave band data, and carrying out pairwise ratio operation on each wave band to construct a characteristic factor set. And respectively calculating the correlation coefficient of the Pearson product moment of each characteristic factor and the water depth value, and selecting the characteristic factor with a larger correlation coefficient for modeling. And reading the characteristic factors, selecting the most suitable machine learning algorithm and the hyper-parameter set, and storing the established model. And evaluating the model precision. And inverting the water depth by using the established model. The invention has the advantages that: the method of ratio and logarithm is used for constructing more factors, the Pearson correlation coefficient is used for selecting the factors, the influence of a large amount of redundant data on model construction is eliminated, and better inversion accuracy can be obtained.
Description
Technical Field
The invention relates to the technical field of remote sensing technology application and water depth measurement, in particular to a shallow sea water depth inversion method, a system, equipment and a medium based on multispectral remote sensing images.
Background
The water depth measurement is an important link of underwater topography measurement, is the basis of all ocean economic and management activities, and the shallow sea water depth is an important topographic factor, so that the method has very important significance on economic and military activities such as offshore transportation, shallow sea resource development and utilization, island climbing operations and the like.
The traditional water depth acquisition is mainly completed by on-site measurement (side-scan sonar, multi-beam and the like), the input manpower and material resources are more, the period is long, the data updating is slow, and partial dangerous and sensitive sea areas are difficult to develop.
In the 60 s of the 20 th century, some scholars began to research and use remote sensing images to measure water depth, and with the development of science and technology, research on extracting water depth by satellite multispectral remote sensing data has also been rapidly developed.
The remote sensing depth measurement can detect the sea areas which cannot be directly entered, such as remote geographic positions, severe environmental conditions and the like, has the advantages of safety, economy, rapidness and high efficiency, can update and dynamically monitor in time, has strong achievement situation, and is the next important research direction.
The principle of remote sensing depth measurement is that solar radiation enters a water body through absorption, scattering and reflection of the atmosphere, reaches the water bottom after being influenced by the absorption and scattering of the water body to light, and reaches a sensor after being reflected by the water bottom and being influenced by the water body and the atmosphere. After the influence of the atmosphere and the water body is removed, the light reflected by the water bottom and entering the sensor reflects the underwater topography, and is the information basis of the water depth remote sensing detection.
At present, the construction aspect of the remote sensing sounding model mainly forms a theoretical interpretation model, a semi-theoretical semi-empirical model, a statistical correlation model and other forms.
Many domestic scholars do much work on remote sensing depth measurement research.
Zhang Zhenxing and the like establish a wave band ratio model through IKONOS multispectral data, perform principal component analysis and transformation on the model, and perform quantitative inversion on shallow sea water depth by applying a neural network technology.
Based on the traditional multiband linear regression model, the mansion Kagchang and the like introduce the technologies of data grouping average preprocessing, tide correction, piecewise linear regression, data normalization and the like, so that the improved model is more reasonable and has higher precision.
Cao Bincai, etc. have studied the method of using ICESat-2 (Ice, cloud, and Elevation SateUite-2) laser satellite data and optical remote sensing image to develop active and passive fusion bathymetry.
In the existing method, a plurality of factors are added into some models, so that a large amount of information redundancy is caused, the calculation amount is increased, and the calculation efficiency is reduced; some segmented fitting techniques can improve inversion accuracy theoretically, but are limited by various conditions in practical application, and have poor practical operability and practicability.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a shallow sea water depth inversion method, a system, equipment and a medium. And establishing an analytical relation between the water depth value and the spectral value by using the known water depth data and the spectral value of the corresponding pixel thereof, so as to predict the water depth condition of the unknown area.
Based on each wave band data of the multispectral remote sensing image, pairwise ratio is carried out on each wave band data, logarithm operation is carried out on each wave band, a characteristic factor combination is constructed, a Pearson correlation coefficient is used as a selection index of the characteristic factor, a factor with high correlation with a water depth value is selected as an independent variable, a mathematical analysis relation between the factor and the water depth is constructed through a statistical machine learning method, and the established model is used for inverting the water depth value of an unknown region, so that good inversion accuracy can be obtained.
In order to realize the purpose, the technical scheme adopted by the invention is as follows:
a shallow sea water depth inversion method comprises the following steps:
the method comprises the following steps: and preprocessing the remote sensing image data.
Radiometric calibration, atmospheric correction and geometric correction are carried out on original remote sensing image data, and the images are subjected to land-water separation by combining an improved normalized water body index (MNDWI) method and a threshold value method.
Step two: and preprocessing the water depth data.
And inquiring the tide table, carrying out tide correction on the water depth data, and correcting the water depth data to be consistent with the image shooting day.
And removing obviously unsuitable water depth points by combining the remote sensing image.
Projective point extraction correspondences using water depth data remote sensing image wave band data of space position.
Step three: and constructing and selecting characteristic factors.
And taking logarithm of each wave band value of the extracted remote sensing image wave band data, and carrying out pairwise ratio operation on each wave band to construct a characteristic factor set.
And respectively calculating a Pearson product-moment correlation coefficient (Pearson product-displacement correlation coefficient) of each characteristic factor and the water depth value, and selecting the characteristic factor with a larger correlation coefficient for modeling.
Step four: and establishing a machine learning model.
Selecting proper program language and machine learning algorithm, reading the characteristic factors processed in the third step, selecting the most suitable machine learning algorithm and hyper-parameter set by using the method of grid search and cross validation, and storing the established model.
Step five: and evaluating the model precision.
The model accuracy was evaluated using three indices, the coefficient of determination (R2), the Root Mean Square Error (RMSE), and the mean relative error (ARE).
Step six: and inverting the water depth by using the established model.
And reading data of each wave band of the remote sensing image, constructing corresponding characteristic factors according to the characteristic factors processed in the third step, and calculating the water depth value of the remote sensing data by using the model established and stored in the fourth step.
Further, in the first step, the MNDWI method extracts the boundary of the water body and the land by utilizing the principle that the reflection of the water body is gradually weakened from visible light to infrared wave bands and the absorption is strongest in the near infrared and middle infrared wavelength ranges.
The MNDWI method formula is as follows:
MNDWI＝(Band Green-Band MIR)/(Band Green+Band MIR)
in the formula, MNDWI is a normalized water body index, bandGreen is the reflectivity of a green light wave band, and BandMIR is the reflectivity of a middle infrared wave band.
Firstly, the whole image is subjected to wave band operation to obtain the MNDWI value of each pixel, and different ground objects are subjected to brightness division.
And extracting the water body by reasonably setting a threshold, reducing the influence of non-water body factors on the water body by setting the threshold to be 0, and dividing all other ground objects outside the water body into background factors so as to obtain a complete water body boundary.
And (4) carrying out binarization on the image, and extracting a vector boundary from the binarized image so as to achieve the purpose of automatically extracting the boundary of the coastal zone.
Further, in step two, the unsuitable water depth points include: data on reefs, buildings and large areas of spoondrifts.
The remote sensing image wave band data comprises: water depth values and 4 band data.
Further, in step three, the natural logarithm is taken for the 4 band data to construct the characteristic factor.
And (4) carrying out pairwise ratio operation on the 4 wave bands respectively to construct a characteristic factor set.
And respectively calculating the Pearson correlation coefficient of the characteristic factor and the water depth value, and selecting the characteristic factor with the correlation coefficient larger than 0.5 for modeling.
Further, the formula used in step five is as follows:
determining a coefficient:
root mean square error:
average relative error:
wherein yi is the actual water depth value of the ith point, yi _ pre is the predicted water depth value of the ith point, y _ avr is the average value of the actual water depth value set, and n is the number of samples. The better the R2 value is closer to 1, the closer the RMSE and ARE ARE to 0, the better the inversion effect is.
The invention also discloses a shallow sea water depth inversion system, which can be used for implementing the shallow sea water depth inversion method, and specifically comprises the following steps: the system comprises a preprocessing module, a characteristic factor constructing and selecting module, a machine learning module, a precision evaluating module and an inversion water depth module;
a preprocessing module: carrying out radiometric calibration, atmospheric correction and geometric correction on original remote sensing image data, and carrying out land and water separation on the image by combining an improved normalized water body index method and a threshold value method.
And inquiring a tide table, carrying out tide correction on the water depth data, correcting the water depth data to be consistent with the current day of image shooting, eliminating obviously unsuitable water depth points by combining the remote sensing image, and extracting the remote sensing image wave band data corresponding to the space position by using the projection points of the water depth data.
A characteristic factor construction and selection module: and taking logarithm of each wave band value of the extracted remote sensing image wave band data, and carrying out pairwise ratio operation on each wave band to construct a characteristic factor set.
And respectively calculating the correlation coefficient of the Pearson product moment of each characteristic factor and the water depth value, and selecting the characteristic factor with a larger correlation coefficient for modeling.
A machine learning module: selecting a proper program language and a machine learning algorithm, reading the characteristic factors processed by the characteristic factor constructing and selecting module, selecting the most suitable machine learning algorithm and the hyper-parameter set by using a grid searching and cross validation method, and storing the established model.
And a precision evaluation module: the model accuracy was evaluated using three indices, the coefficient of determination (R2), the Root Mean Square Error (RMSE), and the mean relative error (ARE).
An inversion water depth module: reading data of each wave band of the remote sensing image, constructing corresponding characteristic factors according to the characteristic factors used by the characteristic factor constructing and selecting module, and calculating the water depth value of the remote sensing data by using the machine learning module.
The invention also discloses computer equipment which comprises a memory, a processor and a computer program which is stored on the memory and can run on the processor, wherein the processor executes the program to realize the shallow sea water depth inversion method.
The invention also discloses a computer readable storage medium, on which a computer program is stored, which when executed by a processor implements the shallow sea water depth inversion method.
Compared with the prior art, the invention has the advantages that:
the method of ratio and logarithm is used for constructing more factors, the Pearson correlation coefficient is used for selecting the factors, the influence of a large amount of redundant data on model construction is eliminated, and better inversion accuracy can be obtained.
Drawings
FIG. 1 is a flow chart of a shallow sea water depth inversion method according to an embodiment of the invention;
FIG. 2 is a diagram illustrating correlation coefficients of various factors and water depth in an embodiment of the present invention;
FIG. 3 is a comparison between the predicted value and the actual value of the model training set versus scatter in accordance with an embodiment of the present invention;
FIG. 4 is a comparison of scatter plots of predicted values and actual values of a model test set according to an embodiment of the present invention;
FIG. 5 is a water depth map of the model inversion according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be further described in detail below with reference to the accompanying drawings by way of examples.
The embodiment is based on WorldView-2 remote sensing image data of a certain area and published water depth point data in a chart.
As shown in fig. 1, the method comprises the following steps:
the method comprises the following steps: and preprocessing the remote sensing data.
The remote sensing data is subjected to radiometric calibration and atmospheric correction by using ENVI software, and an original DN value recorded by the remote sensing data is converted into the reflectivity (or called radiance value) of the outer surface of the atmosphere, so that the interference is eliminated, and the data of the real reflectivity is obtained.
And combining Arcgis software with ground control points to carry out geographic registration on the remote sensing image, and ensuring that the water depth data corresponds to the spatial position of the remote sensing image pixel one by one.
Land and water boundaries are extracted using a combination of a modified normalized water body index (MNDWI) method and a threshold method.
The MNDWI method is that the boundary of the water body and the land is extracted by utilizing the principle that the reflection of the water body is gradually weakened from visible light to infrared wave bands and has the strongest absorption in the near infrared and middle infrared wavelength ranges, and the method can effectively reduce the information of other elements such as soil and the like and effectively enhance the river water body information.
The MNDWI method formula is as follows:
MNDWI＝(Band Green-Band MIR)/(Band Green+Band MIR)
in the formula, MNDWI is a normalized water body index, bandGreen is the reflectivity of a green light wave band, and BandMIR is the reflectivity of a middle infrared wave band.
Firstly, the whole image is subjected to wave band operation to obtain the MNDWI value of each pixel, and different ground objects are subjected to brightness division.
And extracting the water body by reasonably setting a threshold, reducing the influence of non-water body factors on the water body by setting the threshold to be 0, and dividing all other ground objects outside the water body into background factors so as to obtain a complete water body boundary.
And (4) carrying out binarization on the image, and extracting a vector boundary from the binarized image so as to achieve the purpose of automatically extracting the boundary of the coastal zone.
Step two: and processing the water depth data.
And inquiring the tide table data of a port near the survey area in the Chinese maritime service network, correcting the water depth data to the water depth of the current day of remote sensing image shooting, and ensuring that the inverted water depth is consistent with the water depth of the current day of remote sensing image shooting.
The water depth data is projected into a point file of ". Shp" by using Arcgis software, and data such as falling on reefs, buildings and large-area spoondrifts are manually removed by combining remote sensing images.
Extracting the data of 4 wave bands of the remote sensing image to point attributes, and deriving the water depth value (H) and the data of 4 wave bands (B1, B2, B3 and B4) into excel table data.
Step three: and constructing and selecting characteristic factors.
And (4) taking the natural logarithm of the data of 4 wave bands (B1, B2, B3 and B4) in the excel table to construct 4 characteristic factors such as Ln _ B1, ln _ B1 and Ln _ B1.
And performing pairwise ratio operation on the 4 wave bands respectively to construct 6 characteristic factor sets such as B1/B2, B1/B3, B1/B4, B2/B3, B2/B4 and B3/B4.
Thus, a total of 14 characteristic factors are obtained.
The pearson correlation coefficients of 14 factors and the water depth value (H) are calculated respectively, and in the embodiment, the pearson correlation coefficients of the factors and the water depth value (H) can be calculated by using a "pearson ()" function in a third-party library "Scipy" by using Python as a programming language.
7 factors with correlation coefficients greater than 0.5 were selected for modeling use.
Step four: and establishing a machine learning model.
In this embodiment, "Python" is selected as a programming language, and a third-party library "sklern" is selected as a main support library for implementing statistical machine learning.
And (4) reading the excel data stored in the step three, and randomly dividing a training set, a verification set and a test set.
In the embodiment, a random forest algorithm (random forest regressor ()) in the 'sklern' is used, a grid search cross validation function (GridSearchCV ()) is combined, an optimum hyper-parameter set is selected, and the established model is stored.
Step five: and evaluating the model precision.
Selecting the common determination coefficient (R2), root Mean Square Error (RMSE) and Average Relative Error (ARE) of the regression problem to the inversion effect of the water depth, and the formula is as follows:
determination coefficient (R2):
root Mean Square Error (RMSE):
average Relative Error (ARE):
wherein yi is the actual water depth value of the ith point, yi _ pre is the predicted water depth value of the ith point, y _ avr is the average value of the actual water depth value set, and n is the number of samples. The better the R2 value is closer to 1, the closer the RMSE and ARE ARE to 0, the better the inversion effect is.
The calculation of each evaluation index is realized by using Python programming according to a formula.
Step six: and inverting the water depth by using the established model.
And programming and reading 4 wave band data of the remote sensing image, and calculating and constructing a data set consisting of 7 characteristic factors by referring to the method in the third step.
And (5) calculating the water depth value of the remote sensing data by using the model stored in the step four.
As shown in fig. 2, according to the implementation in step three, the correlation coefficients between each factor and the water depth are calculated, and after the correlation coefficients are sorted, the factor with better correlation is selected for modeling.
As shown in fig. 3, according to the method of step four, the existing data is used to train and screen out the optimal model, the inversion water depth and the actual water depth of the model training set are compared by a scatter diagram, and the fitting degree of the data training set can be observed.
As shown in fig. 4, according to the method of the sixth step, the remote sensing image data is read, the water depth is predicted, the inversion water depth and the actual water depth of the test set are compared by a scatter diagram, and the data fitting degree of the test set can be observed.
As shown in fig. 5, the water depth prediction result data obtained in the sixth step is subjected to filtering, cutting, data segmentation display and other processing, and geographic map layer elements are superimposed, so that the underwater deformation condition of the research area can be visually shown.
In another embodiment of the present invention, a shallow sea water depth inversion system is provided, which can be used to implement the above shallow sea water depth inversion method, specifically including: the system comprises a preprocessing module, a characteristic factor constructing and selecting module, a machine learning module, a precision evaluating module and an inversion water depth module;
a preprocessing module: carrying out radiometric calibration, atmospheric correction and geometric correction on original remote sensing image data, and carrying out land and water separation on the image by combining an improved normalized water body index method and a threshold value method.
And inquiring a tide table, carrying out tide correction on the water depth data, correcting the water depth data to be consistent with the current day of image shooting, eliminating obviously unsuitable water depth points by combining the remote sensing image, and extracting the remote sensing image wave band data of the corresponding spatial position by using the projection points of the water depth data.
The characteristic factor constructing and selecting module comprises: and taking logarithm of each wave band value of the extracted remote sensing image wave band data, and carrying out pairwise ratio operation on each wave band to construct a characteristic factor set.
And respectively calculating the correlation coefficient of the Pearson product moment of each characteristic factor and the water depth value, and selecting the characteristic factor with a larger correlation coefficient for modeling.
A machine learning module: selecting a proper program language and a machine learning algorithm, reading the characteristic factors processed by the characteristic factor constructing and selecting module, selecting the most suitable machine learning algorithm and the hyper-parameter set by using a grid searching and cross validation method, and storing the established model.
And a precision evaluation module: the model accuracy was evaluated using three indices, the coefficient of determination (R2), the Root Mean Square Error (RMSE), and the mean relative error (ARE).
An inversion water depth module: reading data of each wave band of the remote sensing image, constructing corresponding characteristic factors according to the characteristic factors used by the characteristic factor constructing and selecting module, and calculating the water depth value of the remote sensing data by using the machine learning module.
In yet another embodiment of the present invention, a terminal device is provided that includes a processor and a memory for storing a computer program comprising program instructions, the processor being configured to execute the program instructions stored by the computer storage medium. The Processor may be a Central Processing Unit (CPU), or may be other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable gate array (FPGA) or other Programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, etc., which is a computing core and a control core of the terminal, and is adapted to implement one or more instructions, and is specifically adapted to load and execute one or more instructions to implement a corresponding method flow or a corresponding function; the processor provided by the embodiment of the invention can be used for the operation of the shallow sea water depth inversion method, and comprises the following steps:
the method comprises the following steps: and preprocessing the remote sensing image data.
Radiometric calibration, atmospheric correction and geometric correction are carried out on original remote sensing image data, and the images are subjected to land-water separation by combining an improved normalized water body index (MNDWI) method and a threshold value method.
Step two: and preprocessing the water depth data.
And inquiring the tide table, carrying out tide correction on the water depth data, and correcting the water depth data to be consistent with the image shooting day.
And removing obviously unsuitable water depth points by combining the remote sensing image.
And extracting remote sensing image wave band data corresponding to the space position by using the projection point of the water depth data.
Step three: and constructing and selecting characteristic factors.
And taking logarithm of each wave band value of the extracted remote sensing image wave band data, and carrying out pairwise ratio operation on each wave band to construct a characteristic factor set.
And respectively calculating a Pearson product-moment correlation coefficient (Pearson product-displacement correlation coefficient) of each characteristic factor and the water depth value, and selecting the characteristic factor with a larger correlation coefficient for modeling.
Step four: and establishing a machine learning model.
Selecting proper program language and machine learning algorithm, reading the characteristic factors processed in the third step, selecting the most suitable machine learning algorithm and hyper-parameter set by using the method of grid search and cross validation, and storing the established model.
Step five: and evaluating the model precision.
The model accuracy was evaluated using three indices, the coefficient of determination (R2), the Root Mean Square Error (RMSE), and the mean relative error (ARE).
Step six: and inverting the water depth by using the established model.
And reading data of each wave band of the remote sensing image, constructing corresponding characteristic factors according to the characteristic factors processed in the third step, and calculating the water depth value of the remote sensing data by using the model established and stored in the fourth step.
In still another embodiment of the present invention, the present invention further provides a storage medium, specifically a computer-readable storage medium (Memory), which is a Memory device in a terminal device and is used for storing programs and data. It is understood that the computer readable storage medium herein may include a built-in storage medium in the terminal device, and may also include an extended storage medium supported by the terminal device. The computer-readable storage medium provides a storage space storing an operating system of the terminal. Also, one or more instructions, which may be one or more computer programs (including program code), are stored in the memory space and are adapted to be loaded and executed by the processor. It should be noted that the computer-readable storage medium may be a high-speed RAM memory, or may be a non-volatile memory (non-volatile memory), such as at least one disk memory.
One or more instructions stored in the computer-readable storage medium may be loaded and executed by a processor to implement the corresponding steps of the shallow sea water depth inversion method in the above embodiments; one or more instructions in the computer readable storage medium are loaded by the processor and perform the steps of:
the method comprises the following steps: and preprocessing the remote sensing image data.
Radiometric calibration, atmospheric correction and geometric correction are carried out on original remote sensing image data, and the images are subjected to land-water separation by combining an improved normalized water body index (MNDWI) method and a threshold value method.
Step two: and preprocessing the water depth data.
And inquiring the tide table, carrying out tide correction on the water depth data, and correcting the water depth data to be consistent with the image shooting day.
And removing obviously unsuitable water depth points by combining the remote sensing image.
And extracting remote sensing image wave band data corresponding to the space position by using the projection point of the water depth data.
Step three: and constructing and selecting characteristic factors.
And taking logarithm of each wave band value of the extracted remote sensing image wave band data, and carrying out pairwise ratio operation on each wave band to construct a characteristic factor set.
And respectively calculating a Pearson product-moment correlation coefficient (Pearson product-displacement correlation coefficient) of each characteristic factor and the water depth value, and selecting the characteristic factor with a larger correlation coefficient for modeling.
Step four: and establishing a machine learning model.
Selecting proper program language and machine learning algorithm, reading the characteristic factors processed in the third step, selecting the most suitable machine learning algorithm and hyper-parameter set by using the method of grid search and cross validation, and storing the established model.
Step five: and evaluating the model precision.
The model accuracy was evaluated using three indices, the coefficient of determination (R2), the Root Mean Square Error (RMSE), and the mean relative error (ARE).
Step six: and inverting the water depth by using the established model.
And reading data of each wave band of the remote sensing image, constructing corresponding characteristic factors according to the characteristic factors processed in the third step, and calculating the water depth value of the remote sensing data by using the model established and stored in the fourth step.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It will be appreciated by those of ordinary skill in the art that the examples described herein are intended to assist the reader in understanding the manner in which the invention is practiced, and it is to be understood that the scope of the invention is not limited to such specifically recited statements and examples. Those skilled in the art can make various other specific changes and combinations based on the teachings of the present invention without departing from the spirit of the invention, and these changes and combinations are within the scope of the invention.
Claims (8)
1. A shallow sea water depth inversion method is characterized by comprising the following steps:
the method comprises the following steps: preprocessing the remote sensing image data;
carrying out radiometric calibration, atmospheric correction and geometric correction on original remote sensing image data, and carrying out land and water separation on the image by combining an improved normalized water body index (MNDWI) method and a threshold value method;
step two: preprocessing the water depth data;
inquiring a tide table, carrying out tide correction on the water depth data, and correcting the water depth data to be consistent with the image shooting day;
removing obviously unsuitable water depth points by combining the remote sensing image;
extracting remote sensing image band data of a corresponding space position by using projection points of the water depth data;
step three: constructing and selecting characteristic factors;
taking logarithm of each wave band value of the extracted remote sensing image wave band data, and carrying out pairwise ratio operation on each wave band to construct a characteristic factor set;
respectively calculating a Pearson product-moment correlation coefficient (Pearson product-moment correlation coefficient) of each characteristic factor and the water depth value, and selecting the characteristic factor with a larger correlation coefficient for modeling;
step four: establishing a machine learning model;
selecting a proper program language and a machine learning algorithm, reading the characteristic factors processed in the third step, selecting the most suitable machine learning algorithm and a hyper-parameter set by using a grid search and cross validation method, and storing the established model;
step five: evaluating the model precision;
evaluating the model precision by using three indexes of a decision coefficient (R2), a Root Mean Square Error (RMSE) and an Average Relative Error (ARE);
step six: inverting the water depth by using the established model;
and reading data of each wave band of the remote sensing image, constructing corresponding characteristic factors according to the characteristic factors processed in the third step, and calculating the water depth value of the remote sensing data by using the model established and stored in the fourth step.
2. The shallow sea water depth inversion method according to claim 1, wherein: in the first step, the MNDWI method is used for extracting the boundary of the water body and the land by utilizing the principle that the reflection of the water body is gradually weakened from visible light to an infrared band and the absorption is strongest in the near infrared and mid-infrared wavelength ranges;
the MNDWI method formula is as follows:
MNDWI＝(Band Green-Band MIR)/(Band Green+Band MIR)
in the formula, bandGreen is the reflectivity of a green light wave band, and BandMIR is the reflectivity of a middle infrared wave band;
firstly, performing band operation on the whole image to obtain the MNDWI value of each pixel, and performing brightness division on different ground objects;
extracting the water body by reasonably setting a threshold, reducing the influence of non-water body factors on the water body by setting the threshold to be '0', and dividing all other ground objects except the water body into background factors so as to obtain a complete water body boundary;
and (4) carrying out binarization on the image, and extracting a vector boundary from the binarized image so as to achieve the purpose of automatically extracting the boundary of the coastal zone.
3. The shallow sea water depth inversion method according to claim 1, wherein: in step two, unsuitable water depth points include: data on reefs, buildings and large-area spoondrifts;
the remote sensing image wave band data comprises: water depth values and 4 band data.
4. The shallow sea water depth inversion method according to claim 1, wherein: in the third step, the natural logarithm is taken from the 4 wave band data, and a characteristic factor is constructed;
carrying out pairwise ratio operation on the 4 wave bands respectively to construct a characteristic factor set;
and respectively calculating the Pearson correlation coefficient of the characteristic factor and the water depth value, and selecting the characteristic factor with the correlation coefficient larger than 0.5 for modeling.
5. The shallow sea water depth inversion method according to claim 1, wherein: the formula used in the step five is as follows:
determining a coefficient:
root mean square error:
average relative error:
wherein yi is the actual water depth value of the ith point, yi _ pre is the predicted water depth value of the ith point, y _ avr is the average value of the actual water depth value set, and n is the number of samples; the closer the R2 value is to 1, the better, the closer the RMSE and ARE ARE to 0, the better the inversion effect.
6. A shallow sea depth inversion system, comprising: the system comprises a preprocessing module, a characteristic factor constructing and selecting module, a machine learning module, a precision evaluating module and an inversion water depth module;
a preprocessing module: carrying out radiometric calibration, atmospheric correction and geometric correction on original remote sensing image data, and carrying out land and water separation on the image by combining an improved normalized water body index method and a threshold value method;
inquiring a tide table, carrying out tide correction on the water depth data, correcting the water depth data to be consistent with the current day of image shooting, eliminating obviously unsuitable water depth points by combining the remote sensing image, and extracting remote sensing image wave band data corresponding to a spatial position by using projection points of the water depth data;
a characteristic factor construction and selection module: taking logarithm of each wave band value of the extracted remote sensing image wave band data, and carrying out pairwise ratio operation on each wave band to construct a characteristic factor set;
respectively calculating the correlation coefficient of the Pearson product moment of each characteristic factor and the water depth value, and selecting the characteristic factor with a larger correlation coefficient for modeling;
a machine learning module: selecting a proper program language and a machine learning algorithm, reading the characteristic factors processed by the characteristic factor construction and selection module, selecting the most suitable machine learning algorithm and a hyper-parameter set by using a grid search and cross validation method, and storing the established model;
and a precision evaluation module: evaluating the model precision by using three indexes of a decision coefficient (R2), a Root Mean Square Error (RMSE) and an Average Relative Error (ARE);
an inversion water depth module: reading data of each wave band of the remote sensing image, constructing corresponding characteristic factors according to the characteristic factors used by the characteristic factor constructing and selecting module, and calculating the water depth value of the remote sensing data by using the machine learning module.
7. A computer device, characterized by: comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the method of shallow water depth inversion according to one of claims 1 to 5 when executing the program.
8. A computer-readable storage medium characterized by: a computer program is stored which, when being executed by a processor, carries out the shallow sea depth inversion method according to one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211362348.6A CN115761463A (en) | 2022-11-02 | 2022-11-02 | Shallow sea water depth inversion method, system, equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211362348.6A CN115761463A (en) | 2022-11-02 | 2022-11-02 | Shallow sea water depth inversion method, system, equipment and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115761463A true CN115761463A (en) | 2023-03-07 |
Family
ID=85355361
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211362348.6A Pending CN115761463A (en) | 2022-11-02 | 2022-11-02 | Shallow sea water depth inversion method, system, equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115761463A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117036777A (en) * | 2023-07-04 | 2023-11-10 | 宁波大学 | Mud flat extraction method based on hyperspectral data |
| CN117274831A (en) * | 2023-09-04 | 2023-12-22 | 大连海事大学 | Offshore turbid water body depth inversion method based on machine learning and hyperspectral satellite remote sensing image |
-
2022
- 2022-11-02 CN CN202211362348.6A patent/CN115761463A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117036777A (en) * | 2023-07-04 | 2023-11-10 | 宁波大学 | Mud flat extraction method based on hyperspectral data |
| CN117274831A (en) * | 2023-09-04 | 2023-12-22 | 大连海事大学 | Offshore turbid water body depth inversion method based on machine learning and hyperspectral satellite remote sensing image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113095303B (en) | Model training method, forest land change detection system, forest land change detection device and forest land change detection medium | |
| CN109934200B (en) | RGB color remote sensing image cloud detection method and system based on improved M-Net | |
| US11521379B1 (en) | Method for flood disaster monitoring and disaster analysis based on vision transformer | |
| CN113705478B (en) | Mangrove single wood target detection method based on improved YOLOv5 | |
| KR20220000898A (en) | Method to identify shoreline changes based on multi-factor | |
| Naboureh et al. | An integrated object-based image analysis and CA-Markov model approach for modeling land use/land cover trends in the Sarab plain | |
| CN103353988B (en) | Allos SAR scene Feature Correspondence Algorithm performance estimating method | |
| CN103063202A (en) | Cyanobacteria biomass spatial-temporal change monitoring and visualization method based on remote sensing image | |
| CN112347895A (en) | Ship remote sensing target detection method based on boundary optimization neural network | |
| Li et al. | A comparison of deep learning methods for airborne lidar point clouds classification | |
| CN115452759B (en) | River and lake health index evaluation method and system based on satellite remote sensing data | |
| Singh et al. | Review on different change vector analysis algorithms based change detection techniques | |
| Li et al. | An iterative-mode scan design of terrestrial laser scanning in forests for minimizing occlusion effects | |
| CN115761463A (en) | Shallow sea water depth inversion method, system, equipment and medium | |
| CN116071664A (en) | SAR image ship detection method based on improved CenterNet network | |
| CN108765439A (en) | A kind of sea horizon detection method based on unmanned water surface ship | |
| CN116295285A (en) | Shallow sea water depth remote sensing inversion method based on region self-adaption | |
| Zhang et al. | A multiband model with successive projections algorithm for bathymetry estimation based on remotely sensed hyperspectral data in Qinghai Lake | |
| CN113902792A (en) | Building height detection method and system based on improved RetinaNet network and electronic equipment | |
| CN117115666B (en) | Plateau lake extraction method, device, equipment and medium based on multi-source data | |
| Oishi et al. | Preliminary verification for application of a support vector machine-based cloud detection method to GOSAT-2 CAI-2 | |
| Ayushi et al. | A comparative analysis of machine learning techniques for aboveground biomass estimation: A case study of the Western Ghats, India | |
| US20210383546A1 (en) | Learning device, image processing device, learning method, image processing method, learning program, and image processing program | |
| CN116263735A (en) | Robustness assessment method, device, equipment and storage medium for neural network | |
| Ilham et al. | Machine learning-based mangrove land classification on worldview-2 satellite image in nusa Lembongan Island |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |