CN116704331A

CN116704331A - Beach change information extraction method and device and electronic equipment

Info

Publication number: CN116704331A
Application number: CN202310538815.4A
Authority: CN
Inventors: 邱玉宝; 梁怡邦
Original assignee: Aerospace Information Research Institute of CAS
Current assignee: Aerospace Information Research Institute of CAS
Priority date: 2023-05-12
Filing date: 2023-05-12
Publication date: 2023-09-05

Abstract

The invention provides a beach change information extraction method, a beach change information extraction device and electronic equipment, and relates to the technical field of remote sensing image information extraction, wherein the method comprises the following steps: acquiring a remote sensing image sequence acquired at a target time resolution in a target period corresponding to a target region; respectively obtaining first surface water vectors corresponding to the first remote sensing images based on the first remote sensing images in the remote sensing image sequences; obtaining second surface water vectors corresponding to the first surface water vectors respectively by using a region growing algorithm based on the first surface water vectors and the first remote sensing images corresponding to the first surface water vectors respectively; performing topological relation operation on the first surface water body vector and the second surface water body vector to obtain beach vectors respectively corresponding to the first remote sensing images; and determining the mud flat change information corresponding to the target area based on each mud flat vector. The method can realize high space-time precision accurate extraction of the beach change information of the target area.

Description

Beach change information extraction method and device and electronic equipment

Technical Field

The invention relates to the technical field of remote sensing image information extraction, in particular to a beach change information extraction method and device and electronic equipment.

Background

The beach is a general term for beach, river beach and lake beach, is called as "intertidal zone" in geomorphology, and is an important land resource and space resource. Therefore, the method has great significance in monitoring the dynamic change of the beach.

Because the remote sensing technology has the characteristics of macroscopic, rapid, high-frequency observation, low cost and the like, the remote sensing technology can be used for monitoring various earth surface objects and land coverage types in a large range, and therefore, the remote sensing technology becomes an effective method for monitoring dynamic changes of the beach in recent years. However, in the dynamic change monitoring of the beach, due to the common influence of the imaging mode, the atmospheric condition, the tide and the like, it is difficult for the related technology to accurately extract the change information of the beach based on the obtained remote sensing image.

Therefore, how to accurately extract the beach change information with high accuracy is a problem that needs to be solved in the industry.

Disclosure of Invention

Aiming at the problems existing in the prior art, the invention provides a method and a device for extracting beach change information and electronic equipment.

In a first aspect, the present invention provides a method for extracting beach change information, including:

acquiring a remote sensing image sequence acquired at a target time resolution in a target period corresponding to a target region;

Respectively obtaining first surface water vectors corresponding to the first remote sensing images based on the first remote sensing images in the remote sensing image sequence;

obtaining second surface water vectors corresponding to the first surface water vectors respectively by using a region growing algorithm based on the first surface water vectors and the first remote sensing images corresponding to the first surface water vectors respectively;

performing topological relation operation on the first surface water body vector and the second surface water body vector which correspond to the first remote sensing images respectively to obtain beach vectors which correspond to the first remote sensing images respectively;

and determining the mud flat change information corresponding to the target area based on each mud flat vector.

Optionally, according to the method for extracting beach change information provided by the present invention, the obtaining a first surface water body vector corresponding to each first remote sensing image based on each first remote sensing image in the remote sensing image sequence respectively includes:

respectively obtaining polarization wave bands, topographic features and texture features corresponding to the first remote sensing images based on the first remote sensing images;

Constructing multidimensional feature spaces corresponding to the first remote sensing images respectively based on the polarization wave bands, the topographic features and the texture features corresponding to the first remote sensing images respectively;

and respectively obtaining first surface water vectors corresponding to the first remote sensing images based on the multidimensional feature spaces.

Optionally, according to the method for extracting beach change information provided by the present invention, the obtaining a first surface water body vector corresponding to each of the first remote sensing images based on each of the multidimensional feature spaces includes:

and respectively inputting the multidimensional feature spaces into a support vector machine classification model to obtain first surface water vectors which are output by the support vector machine classification model and respectively correspond to the first remote sensing images.

Optionally, according to the method for extracting beach change information provided by the present invention, after each of the multidimensional feature spaces is input into a support vector machine classification model, a first surface water body vector corresponding to each of the first remote sensing images output by the support vector machine classification model is obtained, and the method further includes:

and denoising the first surface water body vector to obtain the denoised first surface water body vector.

Optionally, according to the method for extracting beach change information provided by the present invention, before each of the multidimensional feature spaces is input into a support vector machine classification model to obtain a first surface water body vector corresponding to each of the first remote sensing images output by the support vector machine classification model, the method further includes:

determining a segmentation threshold for segmenting the body of water from the non-body of water;

based on the segmentation threshold, respectively carrying out threshold segmentation processing on each sample remote sensing image to obtain target vector data respectively corresponding to each sample remote sensing image, wherein the target vector data comprises water vector data and non-water vector data;

respectively selecting a plurality of sample points from each target vector data to obtain a plurality of sample point sets;

training an initial support vector machine classification model based on the plurality of sample point sets and a cross-checking method to obtain the trained support vector machine classification model.

Optionally, according to the method for extracting beach change information provided by the present invention, the obtaining, by using a region growing algorithm, a second surface water vector corresponding to each of the first remote sensing images based on each of the first surface water vectors and the first remote sensing images corresponding to each of the first surface water vectors, respectively, includes:

Respectively determining target remote sensing images corresponding to VH polarization bands in the polarization bands corresponding to the first remote sensing images;

respectively taking each first surface water body vector as a mask, extracting pixels in each target remote sensing image as a seed point set, and respectively determining threshold ranges which respectively correspond to each seed point set and meet the growth conditions based on each seed point set;

and executing region growing based on each target remote sensing image, each seed point set and each threshold range, and obtaining second surface water vectors corresponding to each first remote sensing image under the condition that each seed point set is determined to be empty.

Optionally, according to the method for extracting beach change information provided by the present invention, the determining, based on each of the seed point sets, a threshold range that meets a growth condition and corresponds to each of the seed point sets, includes:

respectively determining an average value and a standard deviation value of seed points in each seed point set;

and determining threshold ranges which respectively correspond to the seed point sets and meet the growth conditions based on the average value and the standard deviation value respectively corresponding to the seed point sets.

Optionally, according to the method for extracting beach change information provided by the present invention, the obtaining, based on each of the first remote sensing images, a polarization band, a topography feature and a texture feature corresponding to each of the first remote sensing images respectively includes:

preprocessing each first remote sensing image to obtain a polarized wave band corresponding to each first remote sensing image and a second remote sensing image corresponding to the polarized wave band;

based on the second remote sensing images, obtaining the topographic features and the textural features respectively corresponding to the first remote sensing images;

wherein the preprocessing includes thermal noise removal, orbit correction, radiation calibration, speckle filtering, terrain correction, and decibelization processing.

In a second aspect, the present invention further provides a mud flat change information extraction device, including:

the acquisition module is used for acquiring a remote sensing image sequence acquired at a target time resolution in a target period corresponding to the target area;

the first obtaining module is used for obtaining first surface water vectors corresponding to the first remote sensing images respectively based on the first remote sensing images in the remote sensing image sequence respectively;

The second obtaining module is used for obtaining second surface water vectors corresponding to the first surface water vectors respectively by utilizing a region growing algorithm based on the first surface water vectors and the first remote sensing images corresponding to the first surface water vectors respectively;

the third obtaining module is used for respectively carrying out topological relation operation on the first surface water body vector and the second surface water body vector which are respectively corresponding to the first remote sensing images to obtain beach vectors respectively corresponding to the first remote sensing images;

and the determining module is used for determining the beach change information corresponding to the target area based on each beach vector.

In a third aspect, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor implements the beach change information extraction method according to the first aspect when executing the program.

According to the mud flat change information extraction method, the mud flat change information extraction device and the electronic equipment, a remote sensing image sequence acquired at a target time resolution in a target period corresponding to a target area is acquired, then a plurality of first surface water vectors are respectively obtained based on each first remote sensing image in the remote sensing image sequence, further a plurality of second surface water vectors are respectively obtained based on each first surface water vector and the first remote sensing image corresponding to each first surface water vector by using an area growing algorithm, topology relation operation is further carried out on the plurality of first surface water vectors and the plurality of second surface water vectors, a plurality of mud flat vectors corresponding to the target area are obtained, and finally mud flat change information corresponding to the target area is determined based on the plurality of mud flat vectors; according to the invention, the change of the tide level of the target area is considered when the remote sensing image is selected, so that the remote sensing image sequence acquired with the target time resolution in the target period corresponding to the target area is selected for extracting the beach change information, and the second surface water vector is extracted by using the region growing algorithm on the basis of the extracted first surface water vector and the first remote sensing image, so that the extraction precision of the surface water vector is improved, a plurality of beach vectors corresponding to the target area are obtained by carrying out topological relation operation on the first surface water vector and the second surface water vector, and finally the beach change information corresponding to the target area is determined based on the plurality of beach vectors, thereby realizing the accurate extraction of the beach change information of the target area with high space-time precision.

Drawings

In order to more clearly illustrate the invention or the technical solutions of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.

Fig. 1 is a schematic flow chart of a beach change information extraction method provided by the invention;

FIG. 2 is a schematic diagram of an optimal classification line of the support vector machine provided by the invention;

FIG. 3 is a schematic diagram of classification accuracy contours determined according to C and Gamma, provided by the invention;

FIG. 4 is a schematic flow chart of preprocessing a Sentinel-1A GRD image provided by the invention;

FIG. 5 is a schematic flow chart of mud flat vector extraction provided by the invention;

fig. 6 is a schematic structural diagram of the beach change information extracting apparatus provided by the present invention;

fig. 7 is a schematic diagram of the physical structure of the electronic device provided by the present invention.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

It should be noted that, in the description of the present invention, the terms "first," "second," and the like are used for distinguishing between similar objects and not for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged, as appropriate, such that embodiments of the present invention may be implemented in sequences other than those illustrated or described herein, and that the objects identified by "first," "second," etc. are generally of a type, and are not limited to the number of objects, such as the first object may be one or more.

In order to facilitate a clearer understanding of various embodiments of the present invention, some relevant background knowledge is first presented as follows.

The tidal flat refers to the tidal zone between the high tide level and the low tide level of coastal large tide, the tidal flat between the constant water level of rivers and lakes and the flood level, the tidal flat below the flood level of lakes and rivers, and the tidal flat area between the normal water storage level and the maximum flood level of reservoirs and pits. According to the composition of the substances in the beach, the beach can be divided into three types, namely a rock beach, a beach and a mud beach; according to the tide level, width and gradient, the tide level can be divided into three types of high tide beach, medium tide beach and low tide beach. Because of various types of the banks, the effect of water flow, the sand content of the river and other factors, some of the water is flushed, and the beach moves back to land; some stacking effects are strong, and the beach extends to the direction of water; some are more stable, and the beach range is also more stable. As the effects of seasons, climate changes, and the environment continue to increase, the area of some natural and semi-artificial surface waters decreases rapidly in a short period of time, eventually forming an area of beach between normal water storage and low water levels.

The microwave scattering properties of the beach are related to geologic properties (e.g., fine sand, coarse sand, silt, gravel, etc.), surface roughness, water content, etc. In general, the larger the water content and the surface roughness, the stronger the scattering capability of the mud flat on microwaves, so that the same mud flat has different backscattering intensities on synthetic aperture radar (Synthetic Aperture Radar, SAR) images of different periods. Therefore, it is difficult to directly classify and extract the beach ground object on the SAR image, and a certain technology needs to be developed to realize dynamic monitoring of the beach.

However, the tidal flat is periodically submerged by tidal water, and the tidal flat is muddy, has poor accessibility and frequent changes, so that the conventional investigation method is difficult to meet the change monitoring requirement of the highly dynamic environment. Because the remote sensing technology has the characteristics of macroscopic, rapid, high-frequency observation, low cost and the like, the remote sensing technology can be used for monitoring various surface objects and land coverage types in a large range, and provides possibility for long-term large-area sea, river and lake beach resource investigation and timely monitoring.

However, in the dynamic change monitoring of the tidal flat, because of the common influence of the imaging mode, the atmospheric condition, the tide condition and the like, the remote sensing image data still has a plurality of difficulties in application thereof, such as the problems of tide level correction, the difference in acquisition of large-scale tidal flat information, the low acquisition probability of images at the moment of low tide level and the like, so that the change information of the tidal flat is difficult to accurately acquire.

The method and the device for extracting the beach change information and the electronic equipment provided by the invention are exemplarily introduced below with reference to the accompanying drawings.

Fig. 1 is a flow chart of a method for extracting beach change information, which is provided by the invention, as shown in fig. 1, and includes:

step 100, acquiring a remote sensing image sequence acquired with a target time resolution in a target period corresponding to a target area;

step 110, respectively obtaining first surface water vectors corresponding to the first remote sensing images based on the first remote sensing images in the remote sensing image sequence;

step 120, obtaining second surface water vectors corresponding to the first remote sensing images respectively by using a region growing algorithm based on the first surface water vectors and the first remote sensing images corresponding to the first surface water vectors respectively;

step 130, respectively performing topological relation operation on the first surface water body vector and the second surface water body vector which respectively correspond to the first remote sensing images to obtain beach vectors respectively corresponding to the first remote sensing images;

and 140, determining the mud flat change information corresponding to the target area based on each mud flat vector.

It should be noted that, the execution body of the beach change information extraction method provided by the embodiment of the present invention may be an electronic device, a component in the electronic device, an integrated circuit, or a chip. The electronic device may be a mobile electronic device or a non-mobile electronic device. Illustratively, the mobile electronic device may be a mobile phone, a tablet computer, a notebook computer, a palm computer, a wearable device, an Ultra mobile personal computer (Ultra-mobile Personal Computer, UMPC), a netbook or a personal digital assistant (Personal Digital Assistant, PDA), etc., and the non-mobile electronic device may be a server, a network attached storage (Network Attached Storage, NAS), a personal computer (Personal Computer, PC), a Television (Television, TV), a teller machine or a self-service machine, etc., which is not particularly limited by the embodiments of the present invention.

The technical scheme of the embodiment of the invention is described in detail below by taking a computer to execute the method for extracting the beach change information provided by the invention as an example.

Specifically, in order to overcome the defect that in the prior art, change information of a beach is difficult to accurately obtain based on obtained remote sensing images, the invention obtains a plurality of first surface water vectors based on each first remote sensing image in a remote sensing image sequence by obtaining a remote sensing image sequence which is acquired with target time resolution in a target period corresponding to a target area, obtains a plurality of second surface water vectors based on each first surface water vector and each first remote sensing image corresponding to each first surface water vector by utilizing an area growing algorithm, further carries out topological relation operation on the plurality of first surface water vectors and the plurality of second surface water vectors, determines a plurality of beach vectors corresponding to the target area, and finally determines beach change information corresponding to the target area based on the plurality of beach vectors; according to the invention, the change of the tide level of the target area is considered when the remote sensing image is selected, so that the remote sensing image sequence acquired with the target time resolution in the target period corresponding to the target area is selected for extracting the beach change information, and the second surface water vector is extracted by using the region growing algorithm on the basis of the extracted first surface water vector and the first remote sensing image, so that the extraction precision of the surface water vector is improved, a plurality of beach vectors corresponding to the target area are determined by carrying out topological relation operation on the first surface water vector and the second surface water vector, and finally the beach change information corresponding to the target area is determined based on the plurality of beach vectors, thereby realizing the accurate extraction of the beach change information of the target area with high space-time precision.

In the embodiment of the present invention, the target area refers to an area where the beach change information needs to be extracted. For example, river a is a constant water level to the beach between floods.

It should be noted that, the tidal flat in the embodiment of the present invention may be a tidal flat associated with a lake or a reservoir, where the tidal flat associated with a lake or a reservoir without water or a water passing area of a lake without vegetation is formed by flood of lake water in the reservoir and drought of weather.

Optionally, a remote sensing image sequence acquired at a target time resolution in a target period corresponding to the target area may be acquired, where the target period may be adaptively set based on practical applications, and the embodiment of the present invention is not specifically limited thereto, for example, the target period is one year; the target time resolution may range from 5 to 12 days, for example, the target time resolution may be 5 days, 10 days, 12 days, or the like, which is not particularly limited in the embodiment of the present invention.

For example, the target period is one year, the target time resolution is 12 days, then the first remote sensing image corresponding to the target area every 12 days in the one year is acquired, and the finally obtained remote sensing image sequence should include the first remote sensing image of 30 periods.

Alternatively, the first remote sensing image may be a radar image observed by a sentencel-1 satellite.

The Sentinel-1 is an active microwave remote sensing satellite, is provided with a C-band synthetic aperture radar sensor, and can observe the ground and perform radar imaging all the day and all the weather. The Sentinel No. 1 (Sentinel-1) satellite is an earth observation satellite planned by the European space agency (European Space Agency, ESA) Gobedony, and consists of A, B two satellites, wherein the revisitation period of a single satellite is 12 days, the two satellites are complementary, and the revisitation period is 6 days. Sentinel-1 has a total of 4 imaging modes: a stripe Map Mode (SM), an interference Wide Mode (Interferometric Wide Swath, IW), an ultra Wide Mode (EW), a Wave Mode (Wave Mode, WV).

In the embodiment of the invention, a ground range image (Ground Range Detected, GRD) product of a Sentinel-1A IW imaging mode is adopted, the revisitation period is 12 days (the time resolution is 12 days), the spatial resolution is 10 meters, and the monitoring of the change of the surface water body at high frequency can be realized by adopting two polarization modes of VV and VH (wherein VV refers to one co-polarized mode and VH refers to one cross polarized mode).

Optionally, after the remote sensing image sequence is acquired, a first surface water vector corresponding to each first remote sensing image may be acquired based on each first remote sensing image in the remote sensing image sequence.

Optionally, after obtaining the first surface water vectors corresponding to the first remote sensing images respectively, for each first surface water vector, a second surface water vector corresponding to the first remote sensing image may be obtained by using a region growing algorithm based on the first surface water vector and the first remote sensing image corresponding to the first surface water vector.

In the process of extracting the surface water body vector, the prior art is limited by the problem of selecting sample points, so that the extracted surface water body contour has deviation from the actual situation. In the embodiment of the invention, the second surface water body vector is extracted by utilizing a region growing algorithm on the basis of the first surface water body vector and the first remote sensing image; compared with the image gray threshold segmentation technology, the region growing considers the connectivity of image pixel space, and can improve the extraction precision of the surface water body vector.

Alternatively, after obtaining the second surface water vectors respectively corresponding to the first remote sensing images, the beach vectors respectively corresponding to the first remote sensing images may be obtained by respectively performing topological relation operation on the first surface water vectors and the second surface water vectors respectively corresponding to the first remote sensing images.

Optionally, after the mud flat vectors corresponding to the first remote sensing images are obtained, mud flat change information corresponding to the target area can be determined based on the obtained mud flat vectors of different periods of the target area, so that dynamic change monitoring of mud flat of the target area is realized.

The invention provides a beach change information extraction method, which comprises the steps of acquiring a remote sensing image sequence acquired at a target time resolution in a target period corresponding to a target area, then respectively acquiring a plurality of first surface water vectors based on each first remote sensing image in the remote sensing image sequence, further respectively acquiring a plurality of second surface water vectors based on each first surface water vector and the first remote sensing image corresponding to each first surface water vector by using an area growing algorithm, further carrying out topological relation operation on the plurality of first surface water vectors and the plurality of second surface water vectors to acquire a plurality of beach vectors corresponding to the target area, and finally determining beach change information corresponding to the target area based on the plurality of beach vectors; according to the invention, the change of the tide level of the target area is considered when the remote sensing image is selected, so that the remote sensing image sequence acquired with the target time resolution in the target period corresponding to the target area is selected for extracting the beach change information, and the second surface water vector is extracted by using the region growing algorithm on the basis of the extracted first surface water vector and the first remote sensing image, so that the extraction precision of the surface water vector is improved, a plurality of beach vectors corresponding to the target area are obtained by carrying out topological relation operation on the first surface water vector and the second surface water vector, and finally the beach change information corresponding to the target area is determined based on the plurality of beach vectors, thereby realizing the accurate extraction of the beach change information of the target area with high space-time precision.

Optionally, the obtaining a first surface water body vector corresponding to each first remote sensing image based on each first remote sensing image in the remote sensing image sequence respectively includes:

Specifically, in the embodiment of the present invention, in order to obtain the first surface water vectors respectively corresponding to the first remote sensing images, the polarization bands, the topography features and the texture features respectively corresponding to the first remote sensing images may be obtained first based on the first remote sensing images, and then the multidimensional feature spaces respectively corresponding to the first remote sensing images may be constructed based on the polarization bands, the topography features and the texture features respectively corresponding to the first remote sensing images, and then the first surface water vectors respectively corresponding to the first remote sensing images may be obtained based on the multidimensional feature spaces.

The electromagnetic waves emitted cannot reach the back of the mountain due to the influence of the radar side view imaging mode, and a mountain shadow is formed on the SAR image (first remote sensing image). And because the electromagnetic wave is subjected to specular reflection on the water surface, the echo power received by the water surface is very little, and the backward scattering strength of the water body is similar to that of the mountain shadow, so that the water body and the mountain shadow are difficult to separate through single polarized wave band information. But the position where the mountain shadow is located has obvious topographic features, wherein the gradient is an index that is commonly used to characterize the topographic features. For example, the surface of a body of water tends to be calm, with a slope of 0, and where mountain shadows occur, there are larger slope values.

Therefore, in the embodiment of the invention, in order to eliminate the influence of mountain shadows on the water body extraction result, the digital elevation model (Digital Elevation Model, DEM) and the terrain features such as gradient corresponding to the SAR image are introduced.

Moreover, as the features such as paddy fields, roads and the like in plain areas show weaker backward scattering characteristics on SAR images, the features are difficult to distinguish from water bodies on gray values, and finally the water body extraction precision can be influenced. However, in general, the paddy field is used as cultivated land, most of the paddy fields are in a regular grid shape, the surface presents obvious rules and roughness, and the surface of the water body is smooth and the boundary is irregular. Texture is a fundamental feature of an image, and can express the surface of a ground object and the structural attribute thereof, such as the roughness of the surface, the uniformity of the structure and the like.

Therefore, in the embodiment of the invention, in order to more intuitively distinguish different ground objects such as a water body, a paddy field, a road and the like in the SAR image, three texture features are introduced, which are respectively as follows:

the angular second moment (Angular Second Moment, ASM), reflecting the degree of both uniform image gray level distribution and texture thickness, the more uniform the local, the larger the value thereof;

entropy (ENT), reflecting the randomness of the information quantity and the pel value of the image, when the local change is larger, the corresponding Entropy value is larger;

homogeneity (HOM), which reflects the Homogeneity of the image texture, is greater for different regions, and therefore greater Homogeneity of the water body.

Therefore, based on the polarization wave band, the topographic feature and the texture feature corresponding to the first remote sensing image, the constructed multidimensional feature space comprises a VV polarization wave band, a VH polarization wave band, a DEM, a gradient, a slope direction, an angular second moment, entropy and uniformity.

Alternatively, after the multidimensional feature spaces corresponding to the respective first remote sensing images are constructed, the first surface water vectors corresponding to the respective first remote sensing images may be obtained based on the respective multidimensional feature spaces, respectively.

It can be appreciated that by constructing the multidimensional feature space and further obtaining the first surface water body vector corresponding to the first remote sensing image based on the constructed multidimensional feature space, the embodiment of the invention can eliminate the interference of land features such as paddy fields and mountain shadows and improve the extraction precision of the surface water body under complex terrains and cloud and rain environments.

Optionally, the obtaining a first surface water body vector corresponding to each first remote sensing image based on each multidimensional feature space includes:

and respectively inputting the multidimensional feature spaces into a support vector machine (Support Vector Machine, SVM) classification model to obtain first surface water vectors which are output by the support vector machine classification model and respectively correspond to the first remote sensing images.

Specifically, in the embodiment of the invention, the ground object classification can be performed on each multidimensional feature space based on the support vector machine classification model, so as to obtain the first surface water body vector corresponding to each first remote sensing image.

It should be noted that, since the support vector machine belongs to the supervised classification method, and the supervised classification is based on the pixel, some small-area image spots (noise) are inevitably generated in the classification result, so that from the practical application point of view, the small image spots need to be removed or reclassified.

Optionally, after each multidimensional feature space is respectively input into a support vector machine classification model to obtain first surface water body vectors respectively corresponding to each first remote sensing image and output by the support vector machine classification model, the method further includes:

Optionally, in the embodiment of the present invention, the classification result output by the support vector machine classification model may be processed by methods including, but not limited to, majority/minimality analysis, clustering processing, filtering processing, and the like, to remove salt and pepper noise, so as to obtain first surface water vectors corresponding to each first remote sensing image respectively.

It should be noted that, the Majority analysis adopts a method similar to convolution filtering to classify false pixels in a larger class into the class, and firstly defines a transformation kernel size, and replaces the class of the center pixel with the dominant (most pixel) pixel class in the transformation kernel. And for the accuracy analysis it replaces the class of the center pixel with the class of the secondarily located pixels in the transform kernel.

Preferably, in the embodiment of the present invention, denoising the first surface water vectors corresponding to each first remote sensing image output by the support vector machine classification model by using a Majority analysis method, so as to obtain denoised first surface water vectors.

It can be appreciated that the denoising processing is performed on the first surface water body vector, so that the subsequent denoising processing is beneficial to more accurately obtaining the tidal flat vector corresponding to the target area by using the denoised first surface water body vector.

Optionally, before each multidimensional feature space is respectively input into a support vector machine classification model to obtain first surface water body vectors respectively corresponding to each first remote sensing image and output by the support vector machine classification model, the method further includes:

Specifically, in the embodiment of the invention, before classifying each multidimensional feature space by using a support vector machine classification model, the support vector machine classification model needs to be constructed and trained, specifically, firstly, a segmentation threshold value for segmenting water bodies and non-water bodies is determined, then, based on the segmentation threshold value, threshold segmentation processing is respectively carried out on each sample remote sensing image, target vector data respectively corresponding to each sample remote sensing image is obtained, the target vector data comprises water vector data and non-water vector data, a plurality of sample points are selected from each target vector data respectively, a plurality of sample point sets are obtained, finally, an initial support vector machine classification model is trained based on the obtained plurality of sample point sets and a cross checking method, and finally, the trained support vector machine classification model is obtained.

It should be noted that, the SVM is an optimal boundary classification method based on VC dimension (Vapnik-Chervonenkis Dimension) theory and structural risk minimization criteria, and is a maximum interval classifier defined on a feature space. The purpose of the SVM is to find a compromise value of minimum experience risk and confidence interval, and meanwhile training errors and model complexity are considered, so that a good classification effect can be obtained when the number of samples is small. That is, the SVM maximizes the dataset spacing by finding support vectors, thereby finding the optimal segmentation hyperplane.

FIG. 2 is a schematic diagram of an optimal classification line of the support vector machine according to the present invention, as shown in FIG. 2, showing an optimal classification line in a linearly separable sample, H representing the maximum classification interval between data sets, called optimal hyperplane, H ₁ And H ₂ Two parallel hyperplanes on two sides of the optimal hyperplane H are used for carrying out normalization processing on the sample set, and the maximum classification interval is thatY= +1 and y= -1 are class labels.

It should be noted that, the traditional SVM classification method is mainly used for extracting the water body and constructing a feature space based on the band data of the SAR image, and the extraction of the water body is realized by selecting a sample point training model. However, due to the complex scene of SAR images, it is difficult to accurately describe the difference between water and non-water only with one type of features.

Therefore, the embodiment of the invention considers the characteristic space of the sample constructed by combining a plurality of auxiliary data, maps the nonlinear sample to a high-dimensional space, constructs an optimal classification hyperplane and solves the nonlinear problem.

It should be noted that, for the nonlinear separable samples, a kernel function needs to be introduced. The kernel function can map sample data in a low-dimensional space to a high-dimensional space so that the sample is linearly separable, the calculation formula of which is as follows:

wherein K (X, Z) represents a kernel function,for the mapping function +.>Is->And->Inner volume of (A) (I)>Representing a nonlinear feature mapping process from the input space to the high-dimensional space.

Therefore, the process of finding the optimal classification hyperplane can be regarded as a convex quadratic programming problem solving process after adding the constraint condition. By solving this problem, an optimal classification function is obtained, the expression of which is as follows:

wherein, the liquid crystal display device comprises a liquid crystal display device,and b ^* The parameters determined only by the support vector can be used to regulate the optimal classification plane, l representing the dimension and sign representing the sign function.

Optionally, the sample remote sensing image in the embodiment of the invention may be a product of a land coverage type of a european space agency (European Space Agency, ESA) with a spatial resolution of 10 meters, two kinds of raster data of cultivated land and permanent water body may be extracted based on the product, and further the two kinds of raster data are converted into vector data, 1000 points and 500 points are randomly generated in a vector range, and the positions of the random points are regarded as samples of the cultivated land and the water body. And then combining optical images of Google Earth (GE), selecting 500 mountain sample points, and finally constructing three sample point sets including a water body, a cultivated land and mountain shadows, wherein 500 points, 1000 points and 500 points are respectively formed.

It should be noted that, for the SVM classification model, the penalty coefficient C and the Gamma coefficient are two important parameters. C is called a penalty factor (or penalty coefficient), and if the value is larger, the tolerance to misclassification is lower, and a fitting phenomenon (namely, non-water pixels are regarded as water bodies) may occur; conversely, the higher the tolerance to erroneous classification, the lower the accuracy of the classification result. When the remote sensing image processing platforms such as ENVI (Environment for Visualizing Images) and the like operate the supervision classification (SVM), the parameters C and Gamma are default values. However, under the classification model corresponding to the default parameters, the accuracy of the final classification result is often not the highest. In order to improve the accuracy of the classification result, the embodiment of the invention determines the optimal C and Gamma parameters by a cross-checking method.

It should be noted that, the cross-checking method essentially adopts a grid search algorithm, that is, different parameters are arranged and combined according to a specified step length within a certain range of parameters, each group of parameter combinations is tested, and the group of parameters with the optimal performance index is taken as the value of the final parameter.

Specifically, the process of determining the optimal C and Gamma parameters by a cross-checking method includes: and (3) carrying out value combination on the parameters C and Gamma within a certain value range according to a certain step length, dividing the training data set into k groups under the combination of different C and Gamma, wherein one group is used as verification data (the accuracy of verification model classification), and the other k-1 groups are used as training data. In the combination of each group (C, gamma), each group of data is taken as a verification data sample in turn, so that k times of calculation are needed, and the average value of the classification accuracy obtained by the k times of cross tests is taken as the average classification accuracy under the group (C, gamma) model. And (3) circulating the calculation to obtain the average classification accuracy of the model under different (C, gamma) combinations, and taking the (C, gamma) combination corresponding to the highest average classification accuracy.

It should be noted that, when the average classification accuracy obtained by different combinations of the C and Gamma parameters is the same, the combination with smaller C value may be selected, because the smaller C is, the larger the fault tolerance of the classification model is, so as to avoid the over-fitting phenomenon. If the C values are also the same, the first (C, gamma) group is taken.

Fig. 3 is a schematic diagram of a contour line of classification accuracy determined according to C and Gamma, as shown in fig. 3, which shows a contour diagram formed by C, gamma coefficients and average classification accuracy, wherein the abscissa represents different values of C, the ordinate represents different values of Gamma coefficients, and the numerals on different contours represent average classification accuracy, for example, the range of values of C and Gamma corresponding to the highest (82%) average classification accuracy can be determined from fig. 3.

Optionally, the obtaining, based on each first surface water vector and the first remote sensing image corresponding to each first surface water vector, a second surface water vector corresponding to each first remote sensing image respectively by using a region growing algorithm includes:

Specifically, in the embodiment of the present invention, in order to obtain the second surface water vectors corresponding to the first remote sensing images respectively by using the region growing algorithm, the target remote sensing images corresponding to VH polarization bands in the polarization bands corresponding to the first remote sensing images may be first determined respectively, then each first surface water vector is used as a mask, the pixels in each target remote sensing image are extracted as seed point sets, the threshold ranges meeting the growing condition corresponding to each seed point set are determined based on each seed point set respectively, region growing is performed based on each seed point set and each threshold range, and the second surface water vectors corresponding to each first remote sensing image are obtained when it is determined that each seed point set is empty.

It should be noted that the basic idea of the region growing algorithm is: selecting one or a group of pixels as seed points; determining characteristics and similarity judgment criteria; starting to grow from the first seed point, firstly judging whether pixels in a seed neighborhood meet similarity conditions, merging the pixels and the seeds into the same region if the pixels meet the similarity conditions, and further growing by adopting a similar method by taking the merged pixels as growing points; the above operation is repeated until no more pixels satisfying the condition are merged into the region.

Thus, before performing region growing, the input image y of the region growing algorithm is first determined _i (i=1, 2,3,., n), seed point set x _j (j＝1,2,3,...,m)，x _j ∈y _i Wherein m is<n, threshold range I E [ P ] meeting growth conditions _min ,P _max ]，P _min Representing the smallest pixel value, P _max Representing the maximum pixel value; the region growing process is then performed.

The flow of the region growing algorithm comprises: from the first seed point x ₁ Initially, it is determined whether the pixel values of its eight neighbors satisfy the growing condition I. If x ₁ Has a pel x in eight neighbors meeting the growth condition I _m+1 Then pixel x _m+1 Merging to seed Point set x _j In which x is again ₁ From x _j And deleting in the collection. At the same time, x is _m+1 Setting 1 in the pixel at the coordinate to judge the next seed point x ₂ Up to x _j E phi, lastAnd obtaining binary raster data.

It should be noted that, under different polarization modes, the echo power of the beach and the water received by the SAR is different, so that the two features have different backscattering intensities on different polarization images. In order to select the input images required by region growth, the embodiment of the invention selects a plurality of pixels on two polarization wave bands of VV and VH respectively for analysis with respect to three lake water bodies and two land features along the coastal beach, so that the gray histogram double peak effect formed by the lake bank beach and the water bodies on the VV polarization wave bands is more obvious, the distinguishing property of the two land features under the VV polarization is higher, and the improvement of the water body extraction precision in SVM classification is facilitated; conversely, the backscattering characteristics of the two features on the VH polarized image are not greatly different, and the region growth based on the water body pixel set is developed on the VH polarized image, so that the coastal beach is extracted to the maximum extent. Therefore, the embodiment of the invention selects the target remote sensing image corresponding to the VH polarization band as the input image with the region growing.

It can be understood that after the post-treatment of classifying and removing the salt and pepper noise of the classifying model of the support vector machine, the first surface water vectors corresponding to the first remote sensing images are obtained, and the first surface water vectors can represent the outline of the surface water to the maximum extent, so that the interference of farmland and mountain shadows is eliminated. The first surface water vector may be used as a mask to extract the pixels on the input image (the target remote sensing image corresponding to the VH polarization band) as a set of seed points.

It should be noted that a specific determination method of the threshold range satisfying the growth condition is described below.

Alternatively, the region growing may be performed based on each target remote sensing image, each set of sub-points, and each threshold range, respectively, and the second surface water body vector corresponding to each first remote sensing image may be obtained if it is determined that each set of sub-points is empty.

Optionally, the determining, based on each of the seed point sets, a threshold range that meets a growth condition and corresponds to each of the seed point sets includes:

Specifically, in the embodiment of the present invention, in order to determine the threshold range satisfying the growth condition, the threshold range satisfying the growth condition corresponding to each of the sub-point sets may be determined based on the average value and the standard deviation value corresponding to each of the sub-point sets.

Wherein the average value of the seed point set is calculated by the following formula:

The standard deviation of the seed point set is calculated by the following formula:

alternatively, in an embodiment of the present invention, the threshold range satisfying the growth condition may be [ mean-3×std, mean+3×std ].

Specifically, starting from the first seed point in the seed point set, judging whether the value of eight adjacent pixels of the seed point set accords with the threshold range, and if no pixel accords with the threshold range, judging the adjacent pixel of the next seed point; if the condition is met, adding the neighborhood image element into the seed point set to serve as a new seed point. And so on until the set of seed points is empty.

Optionally, the obtaining, based on the first remote sensing images, a polarization band, a topography feature, and a texture feature corresponding to the first remote sensing images includes:

Specifically, in the embodiment of the present invention, in order to obtain, based on each first remote sensing image, a polarization band, a topography feature and a texture feature corresponding to each first remote sensing image, each first remote sensing image may be first preprocessed to obtain, based on each first remote sensing image, a polarization band corresponding to each first remote sensing image, and a second remote sensing image corresponding to the polarization band, and further based on the second remote sensing image, a topography feature and a texture feature corresponding to each first remote sensing image are obtained, where preprocessing performed on each first remote sensing image includes thermal noise removal, track correction, radiation calibration, speckle filtering, topography correction and decibelization processing.

For example, fig. 4 is a schematic flow chart of preprocessing a Sentinel-1A GRD image, and as shown in fig. 4, thermal noise removal, track correction, radiation calibration, speckle filtering, topography correction and decibelization are sequentially performed on the Sentinel-1A GRD image, so as to finally obtain a polarization band, a topography feature and a texture feature corresponding to the Sentinel-1A GRD image.

Although thermal noise is caused by the back-scattered energy of the SAR receiver, it is rarely present in the SAR image, but it has a certain influence on the quality of the SAR image in calm water, rivers, and areas with low back-scattered values. Therefore, the embodiment of the invention is beneficial to the subsequent effective extraction of the surface water body vector by carrying out thermal noise removal processing on the first remote sensing image.

It should be noted that, due to the influence of external conditions and internal factors, the satellite's flight attitude is slightly dithered or shifted, which may cause a certain systematic error. Therefore, in order to reduce systematic errors caused by orbital motion, the embodiment of the invention uses precise orbit information data to accurately calibrate the orbit so as to obtain accurate orbit position and velocity information.

It should be noted that, since the radar system does not perform the radiation calibration processing on the Level-1 data, the radiation deviation is large, and therefore, the SAR intensity image needs to be converted into the backscatter coefficient image before the data is used. The radiation calibration is used for generating radar back scattering coefficient values corresponding to ground targets, so that quantitative analysis of radar images is facilitated. In the embodiment of the invention, the image can be calibrated into a Beta-naught product, and the Beta-naught product is used as the input of the next step.

It should be noted that, the enhancement and attenuation of the SAR coherent wave can cause the radar image to form periodic speckle noise, the influence of the coherent noise is far greater than other noise, meanwhile, the existence of the speckle noise can reduce the signal-to-noise ratio of the SAR image, and the image characteristics can disappear when serious. In the embodiment of the invention, in order to keep as much image detail information as possible and restrain the influence of noise on interpretation of the remote sensing image, the remote sensing image is subjected to speckle filtering processing.

Alternatively, the remote sensing image may be subjected to a speckle filtering process using a Lee Sigma filter. The Lee Sigma filter filters image noise by averaging pixels in two Sigma ranges of a central pixel in a filter window based on Gaussian distribution Sigma probability, has good edge holding capacity, can effectively reduce bright and dark spots generated due to water surface fluctuation, and has good noise suppression effect in the process of extracting surface water by using SAR images.

The terrain correction includes geocoding (giving actual coordinate information to the remote sensing image) and a terrain radiation correction process. In SAR imaging, the distance in the SAR image is distorted due to the change of the terrain, the incident angle of the satellite sensor, and the like, so that the SAR image is distorted to a certain extent.

Therefore, in the embodiment of the invention, the distance Doppler algorithm can be utilized to perform terrain correction processing on the remote sensing image, so as to reduce the influence of the terrain distortion on the remote sensing image, and the projection coordinate system of the output image is set as UTM Zone 49/World Geodetic System 1984 (WGS 84: world Geodetic System 1984, which is a coordinate system established for the use of the global positioning system (Global Positioning System, GPS)).

It should be noted that, after the remote sensing image is subjected to the processes including thermal noise removal, orbit correction, radiation calibration, speckle filtering and topography correction, a backscattering coefficient of a linear proportion unit is obtained, and the value of the backscattering coefficient is usually a small positive value (the reason is that the transmission distance is far, and the receiver receives very little energy/power finally).

Therefore, after decibel processing is carried out, the range of the radar backscattering coefficient is approximately in common Gaussian distribution, and the storage bit number of the data is stored as float floating point data from double-precision double-type data, so that visualization and data analysis are convenient.

Alternatively, in an embodiment of the present invention, an SRTM (Shuttle Radar Topography Mission, space shuttle radar topography mission) 1Sec 30 m DEM of the target area may be acquired, and the resolution of the DEM resampled to 10 m, to calculate the topography features based on the DEM data.

Alternatively, in embodiments of the present invention, the computation of texture features may be based on second order probability statistics.

It should be noted that the second order probability statistics are texture values calculated using a gray scale spatial correlation matrix, which is a relative frequency matrix, i.e. the frequency with which pixel values appear in two adjacent processing windows (separated by a specific distance and direction), which shows the number of occurrences of a relationship between a pixel and its specific neighborhood. After the decibel processing is completed, two wave band images of Beta0_VV_db and Beta0_VH_db are output, and quantization gray level processing can be performed on the two wave band images so as to improve the calculation rate.

For example, the step size of the texture feature calculation is set to 1, the window size is set to 5×5, and the directions are set to 0 °, 45 °, 90 °, and 135 °, and the texture features of two different polarization bands of VV and VH are finally obtained.

FIG. 5 is a schematic flow chart of extracting a tidal flat vector, as shown in FIG. 5, after an SAR image of a researched area is obtained, preprocessing the SAR image to obtain a polarized wave band, a topographic feature and a textural feature corresponding to the SAR image, constructing a multidimensional feature space based on the polarized wave band, the topographic feature and the textural feature, classifying the multidimensional feature space by an SVM classification method, classifying and post-processing a classification result to obtain a first surface water vector, further extracting pixels in the VH polarized wave band image by taking the first surface water vector as a mask to serve as a seed point set, determining a threshold condition based on an average value and a standard difference value of seed points in the seed point set, further executing area growth until the seed point set is empty, obtaining a second surface water vector, and finally carrying out topological relation operation on the first surface water vector and the second surface water vector to obtain the tidal flat vector.

In this embodiment of the present invention, performing a topological relation operation on the first surface water body vector and the second surface water body vector may include: intersection analysis is performed on the first surface water body vector and the second surface water body vector, and a face element vector of the increased beach portion is obtained based on the analysis result. This process can be understood as deleting (erasing) parts of the body of water, thereby obtaining the beach vector.

It can be understood that in the embodiment of the present invention, if the time resolution is 12 days, the SAR images acquired every 12 days may be respectively extracted according to the mud flat vector extraction flow shown in fig. 5, so that 30-period mud flat vectors may be obtained in one year, and further mud flat change information of the studied area may be obtained based on the 30-period mud flat vectors.

Optionally, in the embodiment of the invention, based on the SAR image with the time resolution of 12 days and the spatial resolution of 10 meters, a support vector machine classification method is adopted to initially obtain the spatial variation characteristics of the surface water body every 12 days in the target area, further based on the characteristic that the backward scattering characteristics of the surface water body and the coastal beach are not greatly different under the VH polarization, the area growth through the surface water body pixel set is selected on the VH polarization image, then the vector result after the area growth processing and the initially obtained surface water body vector are subjected to topological relation operation, and finally the lake and reservoir coastal beach vector result is obtained, thereby realizing the monitoring of the space-time variation of the lake and reservoir coastal beach in the target area.

The description of the beach change information extraction apparatus provided by the present invention is provided below, and the beach change information extraction apparatus described below and the beach change information extraction method described above may be referred to correspondingly to each other.

Fig. 6 is a schematic structural diagram of a beach change information extracting apparatus according to the present invention, as shown in fig. 6, the apparatus includes: an acquisition module 610, a first acquisition module 620, a second acquisition module 630, a third acquisition module 640, and a determination module 650; wherein:

the acquiring module 610 is configured to acquire a remote sensing image sequence acquired with a target time resolution in a target period corresponding to a target area;

the first obtaining module 620 is configured to obtain a first surface water vector corresponding to each first remote sensing image based on each first remote sensing image in the remote sensing image sequence, respectively;

the second obtaining module 630 is configured to obtain, based on each of the first surface water vectors and the first remote sensing images corresponding to each of the first surface water vectors, a second surface water vector corresponding to each of the first remote sensing images by using a region growing algorithm;

the third obtaining module 640 is configured to perform a topological relation operation on the first surface water vector and the second surface water vector corresponding to each of the first remote sensing images, to obtain a beach vector corresponding to each of the first remote sensing images;

The determining module 650 is configured to determine, based on each of the mud flat vectors, mud flat change information corresponding to the target area.

The invention provides a beach change information extraction device, which is characterized in that a plurality of first surface water vectors are obtained based on each first remote sensing image in a remote sensing image sequence by acquiring a remote sensing image sequence acquired with target time resolution in a target period corresponding to a target area, and then a plurality of second surface water vectors are obtained based on each first surface water vector and each first remote sensing image corresponding to each first surface water vector by utilizing an area growing algorithm, topology relation operation is further carried out on the plurality of first surface water vectors and the plurality of second surface water vectors to obtain a plurality of beach vectors corresponding to the target area, and beach change information corresponding to the target area is finally determined based on the plurality of beach vectors; according to the invention, the change of the tide level of the target area is considered when the remote sensing image is selected, so that the remote sensing image sequence acquired with the target time resolution in the target period corresponding to the target area is selected for extracting the beach change information, and the second surface water vector is extracted by using the region growing algorithm on the basis of the extracted first surface water vector and the first remote sensing image, so that the extraction precision of the surface water vector is improved, a plurality of beach vectors corresponding to the target area are obtained by carrying out topological relation operation on the first surface water vector and the second surface water vector, and finally the beach change information corresponding to the target area is determined based on the plurality of beach vectors, thereby realizing the accurate extraction of the beach change information of the target area with high space-time precision.

It should be noted that, the beach change information extraction apparatus provided by the embodiment of the present invention can implement all the method steps implemented by the beach change information extraction method embodiment, and can achieve the same technical effects, and detailed descriptions of the same parts and beneficial effects as those of the method embodiment in the embodiment are omitted.

Fig. 7 is a schematic physical structure of an electronic device according to the present invention, as shown in fig. 7, the electronic device may include: processor 710, communication interface (Communications Interface) 720, memory 730, and communication bus 740, wherein processor 710, communication interface 720, memory 730 communicate with each other via communication bus 740. Processor 710 may invoke logic instructions in memory 730 to perform the beach change information extraction methods provided by the methods described above, including:

Further, the logic instructions in the memory 730 described above may be implemented in the form of software functional units and may be stored in a computer readable storage medium when sold or used as a stand alone product. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.

In another aspect, the present invention also provides a computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions which, when executed by a computer, enable the computer to perform the method of extracting beach change information provided by the above methods, the method comprising:

In still another aspect, the present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, is implemented to perform the above-provided beach change information extraction methods, the method comprising:

The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.

From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.

Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. The beach change information extraction method is characterized by comprising the following steps of:

2. The method for extracting beach change information according to claim 1, wherein the obtaining a first surface water body vector corresponding to each of the first remote sensing images based on each of the first remote sensing images in the remote sensing image sequence, respectively, comprises:

3. The method for extracting beach change information according to claim 2, wherein the obtaining first surface water vectors corresponding to the first remote sensing images, respectively, based on the multidimensional feature spaces, respectively, comprises:

4. The method for extracting beach change information according to claim 3, wherein after the step of inputting each of the multi-dimensional feature spaces into a support vector machine classification model to obtain a first surface water body vector corresponding to each of the first remote sensing images output by the support vector machine classification model, the method further comprises:

5. The method for extracting beach change information according to claim 3, wherein before the step of inputting each of the multidimensional feature spaces into a support vector machine classification model to obtain a first surface water body vector corresponding to each of the first remote sensing images output by the support vector machine classification model, the method further comprises:

6. The method according to any one of claims 2 to 5, wherein the obtaining, based on each of the first surface water vectors and the first remote sensing images corresponding to each of the first surface water vectors, a second surface water vector corresponding to each of the first remote sensing images using a region growing algorithm, respectively, includes:

7. The method for extracting beach change information according to claim 6, wherein the determining a threshold range satisfying a growth condition corresponding to each of the seed point sets, respectively, based on each of the seed point sets, respectively, comprises:

8. The method for extracting beach change information according to claim 2, wherein the obtaining, based on the respective first remote sensing images, polarization bands, topography features, and texture features respectively corresponding to the respective first remote sensing images includes:

9. The utility model provides a mud flat change information extraction element which characterized in that includes:

10. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the beach change information extraction method as claimed in any one of claims 1 to 8 when the program is executed by the processor.