Mangrove forest information drawing method and mangrove forest information drawing system cooperating with SAR (Synthetic Aperture Radar) and optical remote sensing data

Abstract

The invention discloses a mangrove forest information drawing method and system cooperating with SAR and optical remote sensing data. The mangrove forest information drawing method comprises the steps that an SAR image, an optical image and digital elevation model (DEM) data of a target research area are acquired; based on the SAR image and the optical image, obtaining image fusion features, index features, texture features and polarization features of the target research area ground features; carrying out importance sorting on all the obtained features, and sequentially adding the features into a machine learning model according to a sequence from high to low to carry out distinguishable analysis on ground features; selecting a sample based on the high-resolution image, and training an established machine learning model for mangrove forest extraction based on the sample to determine a mangrove forest extraction model; and extracting mangrove forest information in the target research area based on the mangrove forest extraction model, and screening the mangrove forest area in combination with DEM data so as to complete mapping of the mangrove forest information.

Description

Technical Field

[0001] The present invention relates to the field of satellite remote sensing image information extraction and application technology, and more specifically, to a mangrove information mapping method and system that coordinates SAR and optical remote sensing data. Background Art

[0002] Mangroves are a type of swampy, woody plant community distributed in the intertidal zone between tropical and subtropical sea and land. They have special sea-land characteristics and provide a "natural barrier" for the coast. Mangroves have great ecological, economic and social value, and play an irreplaceable role in maintaining biodiversity, protecting coastal environment, reinforcing embankments, sheltering from wind, protecting banks and preventing siltation, refining coastal water environment, and protecting farmland and villages from natural disasters such as hurricanes and tsunamis. At the same time, mangroves are also indicators of global environmental and climate change. In the past few decades, due to the rapid development of society and coastal economy, the existing mangroves have been under tremendous pressure from human activities and natural disturbances, and the total area of ​​mangroves has continued to decrease. Therefore, there is an urgent need to obtain accurate and timely information on the spatial distribution of mangroves in order to effectively evaluate and protect mangroves.

[0003] However, due to the complexity of mangroves and their growing environment, it is still a challenge to achieve fast and accurate information mapping of mangroves. Collaborative multi-source remote sensing data is an effective way to meet this challenge, and the extraction and selection of multi-features of images is the key to accurately extracting mangrove distribution information through collaborative multi-source remote sensing data.

[0004] Therefore, there is a need for a mangrove information mapping method that coordinates SAR and optical remote sensing data. Summary of the invention

[0005] The present invention proposes a mangrove information mapping method and system that coordinates SAR and optical remote sensing data to solve the problem of how to extract mangroves efficiently and accurately.

[0006] In order to solve the above problems, according to one aspect of the present invention, a method for mapping mangrove information using collaborative SAR and optical remote sensing data is provided, the method comprising:

[0007] Obtain SAR images, optical images and digital elevation model (DEM) data of the target study area;

[0008] Acquire image fusion features, index features, texture features and polarization features of the objects in the target research area based on the SAR images and optical images;

[0009] Sort all acquired features by importance, and add them to the machine learning model in descending order to perform distinguishability analysis of features;

[0010] Selecting samples based on high-resolution images, and training an established machine learning model for mangrove extraction based on the samples to determine a mangrove extraction model;

[0011] The mangrove information in the target study area is extracted based on the mangrove extraction model, and the mangrove area is screened in combination with DEM data, thereby completing the mapping of mangrove information.

[0012] Preferably, the method further comprises:

[0013] Before acquiring the features, the SAR image is subjected to multi-view processing, image filtering, matrix generation, radiation calibration and terrain correction processing; the optical image is subjected to radiation calibration, atmospheric correction and image resampling processing.

[0014] Preferably, when acquiring polarization characteristics, the method constructs a characteristic matrix, including:

[0015] The characteristic matrix of dual-polarization SAR data is:

[0016]

[0017] The characteristic matrix of full polarization SAR data is:

[0018]

[0019] Among them, S VV and S HH represents the same-polarization complex scattering coefficient; S VH and S HV represents the cross-polarization complex scattering coefficient; * is the complex conjugate operator.

[0020] Preferably, the method further comprises:

[0021] All the features obtained are ranked by importance based on the extremely randomized tree algorithm.

[0022] Preferably, the screening of mangrove areas in combination with DEM data comprises:

[0023] Mangrove areas were screened based on elevations between -5m and 10m above mean sea level and slopes less than 10°;

[0024] The slope is calculated using the following methods, including:

[0025]

[0026] [dz / dx]=((c+2f+i)-(a+2d+g)) / (8*x_cellsize),

[0027] [dz / dy]=((g+2h+i)-(a+2b+c)) / (8*y_cellsize),

[0028] Among them, slope_degrees is the slope; ATAN() represents the inverse tangent function; [dz / dx] is the rate of change of pixel e in the x direction; [dz / dy] is the rate of change of pixel e in the y direction; ai is the pixel, e is the center pixel; cellsize is the pixel size.

[0029] According to another aspect of the present invention, there is provided a mangrove information mapping system that coordinates SAR and optical remote sensing data, the system comprising:

[0030] Data acquisition unit, used to acquire SAR images, optical images and digital elevation model DEM data of the target study area;

[0031] A feature acquisition unit, used for acquiring image fusion features, index features, texture features and polarization features of the objects in the target research area based on the SAR images and the optical images;

[0032] The importance analysis unit is used to sort the importance of all acquired features, so as to add the features into the machine learning model in descending order to perform distinguishability analysis of the features;

[0033] A model determination unit, configured to select samples based on high-resolution images, and train an established machine learning model for mangrove extraction based on the samples to determine a mangrove extraction model;

[0034] The mangrove extraction unit is used to extract the mangrove information in the target research area based on the mangrove extraction model, and screen the mangrove area in combination with the DEM data, so as to complete the mapping of the mangrove information.

[0035] Preferably, the system further comprises: a pre-processing unit, configured to:

[0036] Before acquiring the features, the SAR image is subjected to multi-view processing, image filtering, matrix generation, radiation calibration and terrain correction processing; the optical image is subjected to radiation calibration, atmospheric correction and image resampling processing.

[0037] Preferably, the feature acquisition unit, when acquiring the polarization feature, constructs a feature matrix, including:

[0038] The characteristic matrix of dual-polarization SAR data is:

[0039]

[0040] The characteristic matrix of full polarization SAR data is:

[0041]

[0042] Among them, S VV and S HH represents the same-polarization complex scattering coefficient; S VH and S HV represents the cross-polarization complex scattering coefficient; * is the complex conjugate operator.

[0043] Preferably, the importance analysis unit further comprises:

[0044] All the features obtained are ranked by importance based on the extremely randomized tree algorithm.

[0045] Preferably, the mangrove extraction unit, in combination with DEM data, performs screening of mangrove areas, including:

[0046] Mangrove areas were screened based on elevations between -5m and 10m above mean sea level and slopes less than 10°;

[0047] The slope is calculated using the following methods, including:

[0048]

[0049] [dz / dx]=((c+2f+i)-(a+2d+g)) / (8*x_cellsize),

[0050] [dz / dy]=((g+2h+i)-(a+2b+c)) / (8*y_cellsize),

[0051] Among them, slope_degrees is the slope; ATAN() represents the inverse tangent function; [dz / dx] is the rate of change of pixel e in the x direction; [dz / dy] is the rate of change of pixel e in the y direction; ai is the pixel, e is the center pixel; cellsize is the pixel size.

[0052] The present invention provides a method and system for mapping mangrove information by coordinating SAR and optical remote sensing data, including: obtaining SAR images, optical images and digital elevation model (DEM) data of a target study area; obtaining image fusion features, index features, texture features and polarization features of objects in the target study area based on the SAR images and optical images; ranking all acquired features in order of importance, so as to sequentially add the features to a machine learning model in order from high to low to perform distinguishability analysis of objects; selecting samples based on high-resolution images, and training a machine learning model established for mangrove extraction based on the samples to determine a mangrove extraction model; extracting mangrove information in the target study area based on the mangrove extraction model, and screening mangrove areas in combination with DEM data, thereby completing mapping of mangrove information. The present invention effectively improves the accuracy of mangrove information extraction, and can achieve efficient and accurate information mapping of mangroves, so as to effectively evaluate and protect mangroves. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] A more complete understanding of exemplary embodiments of the present invention may be obtained by referring to the following drawings:

[0054] Figure 1 A flowchart of a method 100 for mapping mangrove information using coordinated SAR and optical remote sensing data according to an embodiment of the present invention;

[0055] Figure 2 It is an overall flow chart of the mangrove information extraction process according to an embodiment of the present invention;

[0056] Figure 3 A pseudo color image of a polarimetric SAR image taken according to an embodiment of the present invention (R: VV, G: VH, B: VV / VH);

[0057] Figure 4 A Sentinel-2 optical image taken according to an embodiment of the present invention (R: B4, G: B3, B: B2);

[0058] Figure 5 Schematic diagram of the characteristics after the fusion of optical and VH polarization SAR images according to an embodiment of the present invention (R: B4, G: B3, B: B2);

[0059] Figure 6 A schematic diagram of anisotropy (A) information obtained after polarization decomposition of a dual-polarization SAR image according to an embodiment of the present invention;

[0060] Figure 7 A schematic diagram of the Inundated Mangrove Index (IMFI) after feature extraction after fusion of optical and VH polarimetric SAR images according to an embodiment of the present invention;

[0061] Figure 8 A schematic diagram of the result of ranking the importance of features of the ERT algorithm according to an embodiment of the present invention;

[0062] Fig. 9 A schematic diagram of the mangrove information mapping results of the experimental area according to an embodiment of the present invention;

[0063] Fig.10 Schematic diagram of the structure of a mangrove information mapping system 1000 that coordinates SAR and optical remote sensing data according to an embodiment of the present invention. DETAILED DESCRIPTION

[0064] Now, exemplary embodiments of the present invention are described with reference to the accompanying drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. These embodiments are provided to disclose the present invention in detail and completely and to fully convey the scope of the present invention to those skilled in the art. The terms used in the exemplary embodiments shown in the accompanying drawings are not intended to limit the present invention. In the accompanying drawings, the same units / elements are marked with the same reference numerals.

[0065] Unless otherwise specified, the terms (including technical terms) used herein have the commonly understood meanings to those skilled in the art. In addition, it is understood that the terms defined in commonly used dictionaries should be understood to have the same meanings as those in the context of the relevant fields, and should not be understood as idealized or overly formal meanings.

[0066] Figure 1 FIG. 1 is a flow chart of a method 100 for mapping mangrove forest information using SAR and optical remote sensing data according to an embodiment of the present invention. Figure 1 As shown, the mangrove forest information mapping method based on coordinated SAR and optical remote sensing data provided in the embodiment of the present invention effectively improves the accuracy of mangrove forest information extraction, and can realize efficient and accurate information mapping of mangrove forests, so as to effectively evaluate and protect and manage mangrove forests. The mangrove forest information mapping method 100 based on coordinated SAR and optical remote sensing data provided in the embodiment of the present invention starts from step 101. In step 101, SAR images, optical images and digital elevation model DEM data of the target study area are obtained.

[0067] In step 102, image fusion features, index features, texture features, and polarization features of the objects in the target research area are obtained based on the SAR images and the optical images.

[0068] Preferably, the method further comprises:

[0069] Before acquiring the features, the SAR image is subjected to multi-view processing, image filtering, matrix generation, radiation calibration and terrain correction processing; the optical image is subjected to radiation calibration, atmospheric correction and image resampling processing.

[0070] Preferably, when acquiring polarization characteristics, the method constructs a characteristic matrix, including:

[0071] The characteristic matrix of dual-polarization SAR data is:

[0072]

[0073] The characteristic matrix of full polarization SAR data is:

[0074]

[0075] Among them, S VV and S HH represents the same-polarization complex scattering coefficient; S VH and S HV represents the cross-polarization complex scattering coefficient; * is the complex conjugate operator.

[0076] Combination Figure 2 As shown, in the present invention, SAR images, optical images and SRTM DEM data of the study area are first obtained, and the acquired remote sensing images are preprocessed. In a specific embodiment of the present invention, the SAR image is Cosmo-SkyMed data, and the optical image is Sentinel-2 data, and the basic parameters are shown in Table 1 and Table 2, respectively. Then, the acquired data is preprocessed. Among them, the Cosmo-SkyMed SAR image preprocessing steps include multi-view processing, image filtering, matrix generation and radiation calibration, etc., and the Sentinel-2 image preprocessing steps include radiation calibration, atmospheric correction and image resampling, etc.

[0077] In the present invention, a three-level registration method is adopted for the image registration of optical and SAR images: (1) image feature point detection and feature matching to eliminate the possible rotation and offset between the optical and SAR images and achieve coarse image registration; (2) extract typical feature points on the coarse registration image and extract the gradient intensity information of the SAR and optical images through the gradient operator; (3) finally, normalized cross-correlation is performed based on the gradient intensity to extract the feature points with the same name on the SAR and optical images, and the precise image registration is achieved based on the feature points with the same name.

[0078] Table 1 Basic parameters of Cosmo-SkyMed dual-polarization data in the study area

[0079] parameter Numeric Satellite Type Cosmo-SkyMed Spatial resolution 3m×3m Polarization VV+VH Data acquisition time 20230317

[0080] Table 2 Basic parameters of Sentinel-2 data in the study area

[0081]

[0082]

[0083] In a specific embodiment of the present invention, the pseudo color image of the polarimetric SAR image captured is as follows: Figure 3 As shown in the figure, the optical image of Sentinel-2 is as follows Figure 4 shown.

[0084] Combination Figure 2 As shown, in the present invention, the index features, texture features and polarization features of the objects in the study area are obtained based on the pre-processed SAR and optical images to solve the problem that it is difficult to accurately extract the mangrove distribution information by relying on a single data parameter. Among them, the index features include: parameters obtained by calculating the original optical image and parameters obtained by calculating the features after image fusion. The specific feature parameters are obtained as shown in Table 3 below.

[0085] Table 3. Feature parameters of the study area

[0086]

[0087]

[0088] In a specific embodiment of the present invention, image fusion can be performed by Gram-schmidt Pan Sharpening (GS). GS is a high-fidelity image fusion algorithm that uses a statistical analysis method to optimally match each band involved in the fusion, avoiding the problem of excessive concentration of certain band information in traditional fusion methods, and can better maintain spatial texture information, especially high-fidelity spectral characteristics.

[0089] In the specific implementation of the present invention, the acquired SAR image is dual-polarization data, and a dual-polarization SAR feature matrix should be constructed during polarization decomposition:

[0090]

[0091] When the SAR image is fully polarized data, the fully polarized SAR data feature matrix is ​​constructed as:

[0092]

[0093] Among them, S VV and S HH represents the same-polarization complex scattering coefficient; S VH and S HV represents the cross-polarization complex scattering coefficient; * is the complex conjugate operator.

[0094] Continuing with the above image as an example, in the present invention, the characteristics of the optical and VH polarization SAR images after fusion are shown as follows: Figure 5 As shown in Figure 2, the anisotropy (A) information obtained after polarization decomposition of dual-polarization SAR images is as follows: Figure 6 The submerged mangrove index (IMFI) extracted from the fusion of optical and VH polarimetric SAR images is shown in Figure 7 shown.

[0095] In step 103, all acquired features are ranked in importance, so that the features are sequentially added to the machine learning model in descending order to perform distinguishability analysis of the features.

[0096] Preferably, the method further comprises:

[0097] All the features obtained are ranked by importance based on the extremely randomized tree algorithm.

[0098] Since different feature variables contribute differently in the information extraction process, introducing too many feature variables may lead to a decrease in information extraction accuracy. Therefore, in the present invention, all features are ranked in order of importance and added in descending order to perform feature distinguishability analysis, and feature variables used for information extraction are preferably selected.

[0099] In the present invention, the importance of features is ranked based on the Extremely Randomized Tree (ERT) feature optimization algorithm. The ERT algorithm integrates multiple decision trees for scoring, votes according to the average value of the prediction value of each decision tree, and calculates the contribution of the feature to the branch of each tree to evaluate the importance of the feature. Each tree of the ERT algorithm is calculated based on all samples, ensuring the utilization of the samples. The algorithm randomly selects a subset of features at each node to ensure the difference between each decision tree, thereby introducing more randomness, thereby improving the accuracy and generalization ability of the model, and is suitable for more practical application scenarios. The ERT algorithm usually uses the maximum information gain (Information Gain) as the evaluation criterion, and the calculation formula is as follows:

[0100]

[0101]

[0102] In the formula, c refers to the number of sample label categories, p i It refers to the proportion of output label i, and A represents a certain feature.

[0103] In the present invention, the ERT algorithm feature importance ranking results are as follows: Figure 8 shown.

[0104] In step 104, samples are selected based on the high-resolution images, and the established machine learning model for mangrove extraction is trained based on the samples to determine a mangrove extraction model.

[0105] In the present invention, samples are selected using Google high-resolution images to construct training sample sets and verification sample sets for subsequent information extraction. In order to facilitate comparative analysis of the extraction results, the objects in the study area are divided into five types: mangroves, other vegetation, paddy fields, buildings, and water bodies. Visual interpretation is used to select sample data for information extraction on Google images, and the ratio of training samples to verification samples is about 7:3. The samples are input into the established machine learning model for mangrove extraction for training to determine the mangrove extraction model.

[0106] In the present invention, the Random Forest (RF) algorithm is used to construct the model to extract mangrove information. RF is a machine learning algorithm based on the Bagging method. It combines multiple decision trees into a forest through the idea of ​​ensemble learning, and then the independent decision trees vote to determine the final classification result. Using multiple decision trees to complete the task together can effectively solve the under-fitting and over-fitting problems of the classification results of a single decision tree, and achieve better accuracy. The final prediction result can be expressed as:

[0107]

[0108] In the formula, H(x) represents the final prediction result, I(h i (X) = Y) represents the characteristic function, h i represents a single decision tree, and Y represents the output variable.

[0109] In the present invention, after the training is completed, the accuracy of the mangrove extraction results needs to be verified. The accuracy of the classification results is evaluated based on the confusion matrix, and the specific evaluation indicators include Overall Accuracy (OA), Producer Accuracy (PA), User Accuracy (UA) and Kappa coefficient.

[0110] In step 105, the mangrove information in the target study area is extracted based on the mangrove extraction model, and the mangrove area is screened in combination with the DEM data, thereby completing the mapping of the mangrove information.

[0111] Preferably, the screening of mangrove areas in combination with DEM data comprises:

[0112] Mangrove areas were screened based on elevations between -5m and 10m above mean sea level and slopes less than 10°;

[0113] The slope is calculated using the following methods, including:

[0114]

[0115] [dz / dx]=((c+2f+i)-(a+2d+g)) / (8*x_cellsize),

[0116] [dz / dy]=((g+2h+i)-(a+2b+c)) / (8*y_cellsize),

[0117] Among them, slope_degrees is the slope; ATAN() represents the inverse tangent function; [dz / dx] is the rate of change of pixel e in the x direction; [dz / dy] is the rate of change of pixel e in the y direction; ai is the pixel, e is the center pixel; cellsize is the pixel size.

[0118] In the present invention, the mangrove information in the target study area is extracted based on the mangrove extraction model, and the classification results are screened using SRTM DEM data. Most mangroves have elevations between -5m and 10m above mean sea level and slopes less than 10°. Therefore, the mangrove extraction results are screened according to this rule to obtain more accurate distribution information.

[0119] In the present invention, the calculation of the slope depends on the rate of change (increment) of the surface in the horizontal (dz / dx) direction and the vertical (dz / dy) direction starting from the center pixel. The basic algorithm used to calculate the slope is:

[0120]

[0121] Slope is usually measured in degrees and is calculated as follows:

[0122]

[0123] The value 57.29578 shown here is the value obtained by truncating the calculation result of 180 / pi.

[0124] The horizontal and vertical increments are determined based on the values ​​of the center pixel e and its eight neighboring pixels. These neighboring pixels are identified using letters a to i, where e represents the pixel for which the slope is currently being calculated.

[0125] The rate of change of pixel e in the x direction is calculated by the following algorithm:

[0126] [dz / dx]=((c+2f+i)-(a+2d+g)) / (8*x_cellsize)

[0127] The rate of change of pixel e in the y direction is calculated using the following algorithm:

[0128] [dz / dy]=((g+2h+i)-(a+2b+c)) / (8*y_cellsize)

[0129] Continuing with the above image as an example, the mangrove information mapping results of the experimental area obtained by mapping according to the method of the present invention are as follows: Fig. 9 shown.

[0130] The present invention obtains multiple characteristic parameters based on the SAR and optical image data of the study area, and then sorts all the characteristics by importance, adds the characteristics in order from high to low to perform ground feature distinguishability analysis, optimizes the characteristic variables for information extraction, and finally uses a machine learning algorithm to obtain mangrove distribution information. The present invention can accurately and timely obtain mangrove spatial distribution information, and can effectively evaluate and protect mangroves.

[0131] Fig.10 FIG. 1 is a schematic diagram of the structure of a mangrove forest information mapping system 1000 based on coordinated SAR and optical remote sensing data according to an embodiment of the present invention. Fig.10 As shown, the mangrove information mapping system 1000 based on collaborative SAR and optical remote sensing data provided in an embodiment of the present invention includes: a data acquisition unit 1001, a feature acquisition unit 1002, an importance analysis unit 1003, a model determination unit 1004 and a mangrove extraction unit 1005.

[0132] Preferably, the data acquisition unit 1001 is used to acquire SAR images, optical images and digital elevation model DEM data of the target research area.

[0133] Preferably, the system further comprises: a pre-processing unit, configured to:

[0134] Before acquiring the features, the SAR image is subjected to multi-view processing, image filtering, matrix generation, radiation calibration and terrain correction processing; the optical image is subjected to radiation calibration, atmospheric correction and image resampling processing.

[0135] Preferably, the feature acquisition unit 1002 is used to acquire image fusion features, index features, texture features and polarization features of the objects in the target research area based on the SAR images and optical images.

[0136] Preferably, the feature acquisition unit 1002 constructs a feature matrix when acquiring polarization features, including:

[0137] The characteristic matrix of dual-polarization SAR data is:

[0138]

[0139] The characteristic matrix of full polarization SAR data is:

[0140]

[0141] Among them, S VV and S HH represents the same-polarization complex scattering coefficient; S VH and S HV represents the cross-polarization complex scattering coefficient; * is the complex conjugate operator.

[0142] Preferably, the importance analysis unit 1003 is used to sort the importance of all acquired features so as to sequentially add the features into the machine learning model in descending order to perform distinguishability analysis of the features.

[0143] Preferably, the importance analysis unit 1003 further includes:

[0144] All the features obtained are ranked by importance based on the extremely randomized tree algorithm.

[0145] Preferably, the model determination unit 1004 is used to select samples based on high-resolution images, and train the established machine learning model for mangrove extraction based on the samples to determine the mangrove extraction model.

[0146] Preferably, the mangrove extraction unit 1005 is used to extract the mangrove information in the target study area based on the mangrove extraction model, and screen the mangrove area in combination with DEM data, thereby completing the mapping of mangrove information.

[0147] Preferably, the mangrove extraction unit 1005, in combination with DEM data, performs screening of mangrove areas, including:

[0148] Mangrove areas were screened based on elevations between -5m and 10m above mean sea level and slopes less than 10°;

[0149] The slope is calculated using the following methods, including:

[0150]

[0151] [dz / dx]=((c+2f+i)-(a+2d+g)) / (8*x_cellsize),

[0152] [dz / dy]=((g+2h+i)-(a+2b+c)) / (8*y_cellsize),

[0153] Among them, slope_degrees is the slope; ATAN() represents the inverse tangent function; [dz / dx] is the rate of change of pixel e in the x direction; [dz / dy] is the rate of change of pixel e in the y direction; ai is the pixel, e is the center pixel; cellsize is the pixel size.

[0154] The mangrove information mapping system 1000 of the embodiment of the present invention in collaboration with SAR and optical remote sensing data corresponds to the mangrove information mapping method 100 of another embodiment of the present invention in collaboration with SAR and optical remote sensing data, which will not be described in detail here.

[0155] The invention has been described above with reference to a few embodiments. However, it is readily apparent to a person skilled in the art that other embodiments than the ones disclosed above are equally within the scope of the invention, as defined by the appended patent claims.

[0156] Generally, all terms used in the claims are to be interpreted according to their ordinary meaning in the technical field, unless explicitly defined otherwise therein. All references to "a / said / the [means, components, etc.]" are to be openly interpreted as at least one instance of said means, components, etc., unless explicitly stated otherwise. The steps of any method disclosed herein do not necessarily have to be performed in the exact order disclosed, unless explicitly stated otherwise.

[0157] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0158] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0159] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0160] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0161] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for mapping mangrove forest information using SAR and optical remote sensing data, characterized in that: The method comprises: Obtain SAR images, optical images and digital elevation model (DEM) data of the target study area; Acquire image fusion features, index features, texture features and polarization features of the objects in the target research area based on the SAR images and optical images; Sort all acquired features by importance, and add them to the machine learning model in descending order to perform distinguishability analysis of features; Selecting samples based on high-resolution images, and training an established machine learning model for mangrove extraction based on the samples to determine a mangrove extraction model; The mangrove information in the target study area is extracted based on the mangrove extraction model, and the mangrove area is screened in combination with DEM data, thereby completing the mapping of mangrove information.

2. The method according to claim 1, characterized in that The method further comprises: Before acquiring the features, the SAR image is subjected to multi-view processing, image filtering, matrix generation, radiation calibration and terrain correction processing; the optical image is subjected to radiation calibration, atmospheric correction and image resampling processing.

3. The method according to claim 1, characterized in that When obtaining polarization characteristics, the method constructs a characteristic matrix, including: The characteristic matrix of dual-polarization SAR data is: The characteristic matrix of full polarization SAR data is: Among them, S VV and S HH represents the same-polarization complex scattering coefficient; S VH and S HV represents the cross-polarization complex scattering coefficient; * is the complex conjugate operator.

4. The method according to claim 1, characterized in that The method further comprises: All the features obtained are ranked by importance based on the extremely randomized tree algorithm.

5. The method according to claim 1, characterized in that The screening of mangrove areas by combining DEM data includes: Mangrove areas were screened based on elevations between -5m and 10m above mean sea level and slopes less than 10°; The slope is calculated using the following methods, including: [dz / dx]=((c+2f+i)-(a+2d+g)) / (8*x_cellsize), [dz / dy]=((g+2h+i)-(a+2b+c)) / (8*y_cellsize), Among them, slope_degrees is the slope; ATAN() represents the inverse tangent function; [dz / dx] is the rate of change of pixel e in the x direction; [dz / dy] is the rate of change of pixel e in the y direction; ai is the pixel, e is the center pixel; cellsize is the pixel size.

6. A mangrove forest information mapping system based on SAR and optical remote sensing data, characterized in that: The system comprises: Data acquisition unit, used to acquire SAR images, optical images and digital elevation model DEM data of the target study area; A feature acquisition unit, used for acquiring image fusion features, index features, texture features and polarization features of the objects in the target research area based on the SAR images and the optical images; The importance analysis unit is used to sort the importance of all acquired features, so as to add the features into the machine learning model in descending order to perform distinguishability analysis of the features; A model determination unit, configured to select samples based on high-resolution images, and train an established machine learning model for mangrove extraction based on the samples to determine a mangrove extraction model; The mangrove extraction unit is used to extract the mangrove information in the target research area based on the mangrove extraction model, and screen the mangrove area in combination with the DEM data, so as to complete the mapping of the mangrove information.

7. The system according to claim 6, characterized in that The system further comprises a pre-processing unit, configured to: Before acquiring the features, the SAR image is subjected to multi-view processing, image filtering, matrix generation, radiation calibration and terrain correction processing; the optical image is subjected to radiation calibration, atmospheric correction and image resampling processing.

8. The system according to claim 6, characterized in that The feature acquisition unit constructs a feature matrix when acquiring polarization features, including: The characteristic matrix of dual-polarization SAR data is: The characteristic matrix of full polarization SAR data is: Among them, S VV and S HH represents the same-polarization complex scattering coefficient; S VH and S HV represents the cross-polarization complex scattering coefficient; * is the complex conjugate operator.

9. The system according to claim 6, characterized in that The importance analysis unit further includes: All the features obtained are ranked by importance based on the extremely randomized tree algorithm.

10. The system according to claim 6, characterized in that The mangrove extraction unit, combined with DEM data, screens the mangrove area, including: Mangrove areas were screened based on elevations between -5m and 10m above mean sea level and slopes less than 10°; The slope is calculated using the following methods, including: [dz / dx]=((c+2f+i)-(a+2d+g)) / (8*x_cellsize), [dz / dy]=((g+2h+i)-(a+2b+c)) / (8*y_cellsize), Among them, slope_degrees is the slope; ATAN() represents the inverse tangent function; [dz / dx] is the rate of change of pixel e in the x direction; [dz / dy] is the rate of change of pixel e in the y direction; ai is the pixel, e is the center pixel; cellsize is the pixel size.

Images