CN111259876A - Radar data water body information extraction method and system based on land surface water body product - Google Patents
Radar data water body information extraction method and system based on land surface water body product Download PDFInfo
- Publication number
- CN111259876A CN111259876A CN202010370050.4A CN202010370050A CN111259876A CN 111259876 A CN111259876 A CN 111259876A CN 202010370050 A CN202010370050 A CN 202010370050A CN 111259876 A CN111259876 A CN 111259876A
- Authority
- CN
- China
- Prior art keywords
- water body
- data
- water
- product
- training
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/182—Network patterns, e.g. roads or rivers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Radar Systems Or Details Thereof (AREA)
- Image Processing (AREA)
Abstract
The invention provides a radar data water body information extraction method and system based on a land surface water body product, wherein the method comprises the steps of obtaining satellite data and the following steps: preprocessing the satellite data to generate backscattering data and calculating a derivative coefficient of the backscattering data; creating a training sample dataset using global water product data, the backscatter data and the derivative coefficients; extracting water body information of the training sample data based on a random forest model to obtain an initial water body extraction result; and carrying out post-treatment on the initial water body extraction result to obtain a final water body product. The method uses the JRC global water product to automatically generate a training data set, screens the data set, then uses a machine learning classification method to automatically extract water information, performs post-processing on an extracted result, and eliminates images of small image spots and mountain shadows, thereby obtaining an accurate water information product.
Description
Technical Field
The invention relates to the technical field of hydraulic simulation, in particular to a radar data water body information extraction method and system based on a land surface water body product.
Background
The method for rapidly acquiring the water body information distribution has great significance in flood situation assessment, ecological environment monitoring, water resource investigation and the like. The remote sensing technology is an important means for rapidly and accurately acquiring water body information. Optical remote sensing is easily affected by weather due to its short wavelength, so the application is limited to a certain extent. In microwave remote sensing, a satellite-borne Synthetic Aperture Radar (SAR) has the characteristics of all-time and all-weather, is not influenced by cloud, rain and fog, can form images at night, and becomes a powerful tool for flood disaster monitoring and lake dynamic monitoring. The Sentinel-1A series satellites launched in 4 months of 2014 by the European Bureau have a single satellite orbit period of 12 days, two satellites of 6 days, a maximum width of 400km and a maximum resolution of 5m, are provided with single polarization (HH/VV) and dual polarization (HH + HV/VV + VH) imaging, free data and convenient acquisition, and can be downloaded by global users.
At present, SAR water body information extraction mainly comprises two methods, namely a method based on threshold segmentation and a method based on classification. The threshold method is mainly applied to a single threshold method, a multi-threshold method, a double peak method, a global threshold method and the like. However, in the large-scale water body information extraction, the water body occupation ratio is relatively low, the threshold value is relatively difficult to take, and due to the image of the radar imaging mechanism, mountain shadow, terrain distortion and wind wave can exist in the image, so that the water body surface is noisy, and the threshold value method is often poor in extraction effect. And identifying a large range of bodies of water based on classification has been characterized as a successful algorithm. However, classification-based algorithms are often supervised methods, and require enough training samples, which are traditionally selected manually, which greatly hinders the automation efficiency of water body information extraction. Currently, the existing data sets, such as MODIS and SRTM derived water body masks, are used to train the model, so as to automatically extract the water body information, and the method has great potential. However, the spatial resolution of MODIS and SRTM is low, which affects the accuracy and effect of water extraction. The global surface water product (with the resolution of 30m and various products) manufactured by the European Union Research center JRC (Joint Research centre) according to the Landsat data of 1984-2018 has the overall precision as high as 99.6 percent and can be used as a training sample for the fine extraction of water information in a classification method.
Wenli Huang et al proposed an automatic extraction method of water body information based on radar satellites in 2018 on Remote Sensing, and the method comprises the following steps: (1) generating a water body label based on an SRTM water body data set SWBD and a composite dynamic surface water range cDSWE, and forming a training sample by using a radar backscattering coefficient, a derivative index and a local incident angle as characteristics and the water body label; (2) randomly selecting part of training samples from the sample data according to the proportion of the water body sample to the non-water body sample; (3) training a random forest by using a randomly sampled training sample, then applying the trained model to all pixels to generate a water body probability map, and classifying the probability map by using high-probability water bodies, medium-probability water bodies, low-probability water bodies and non-water body labels. The method has the advantages that the extraction speed is high, the accuracy is relatively high, but the extraction accuracy is influenced because obvious wrong samples generated due to seasons, time and the like in training samples are not considered; meanwhile, the influence of mountain shadow is not considered, and part of mountain shadow in the water body is extracted.
Disclosure of Invention
In order to solve the technical problems, the invention provides a radar data water body information extraction method and system based on a land surface water body product.
The invention aims to provide a radar data water body information extraction method based on a land surface water body product, which comprises the following steps of acquiring satellite data:
step 1: preprocessing the satellite data to generate backscattering data and calculating a derivative coefficient of the backscattering data;
step 2: creating a training sample dataset using global water product data, the backscatter data and the derivative coefficients;
and step 3: extracting water body information of the training sample data based on a random forest model to obtain an initial water body extraction result;
and 4, step 4: and carrying out post-treatment on the initial water body extraction result to obtain a final water body product.
Preferably, the preprocessing is to perform orbit correction, radiation correction, speckle filtering, multi-vision, terrain correction and db conversion operations on the intensity of the satellite data after the satellite data is subjected to basic processing and then to generate the backscattering data finallyVHAndVV。
in any of the above embodiments, preferably, the derivative coefficient includes a polarization ratioVHrVVThe calculation formula is
。
Preferably in any of the above aspects, the derivation coefficients further include normalized deviation polarization indexNDPIThe calculation formula is
。
Preferably in any of the above aspects, the derivation coefficients further comprise normalized VH indicesNVHIThe calculation formula is
。
In any of the above aspects, preferably, the derived coefficients further include a normalized VV indexNVVIThe calculation formula is
。
In any of the above schemes, preferably, the step 2 includes separately sampling the water body sample and the non-water body sample according to the aboveVVSorting the values, and selecting the quantiles in the water body sample asN1% as the maximum threshold, deleting samples greater than the maximum threshold; the water body sample has the following componentsN2% value as the minimum threshold, samples less than the minimum threshold are deleted.
In any of the above solutions, preferably, the step 3 includes the following sub-steps:
step 31: randomly selecting from training data setsN3% water samples, when the number of the water samples is less than a number threshold, selecting all the water samples;
step 32: determining the optimal parameters of a random forest classifier by using a cross validation grid GridSearch, and generating a training model;
step 33: applying the training model to all pixels of the whole image to obtain the probability that each pixel is a water body;
step 34: and obtaining water body distribution according to the threshold value to obtain an initial water body extraction result.
In any of the above schemes, preferably, the step 4 includes the following sub-steps:
step 41: calculating a gradient map by using a digital elevation model DEM of srtm 30;
step 42: assigning the corresponding extraction result with the gradient larger than the gradient threshold value in the initial water body extraction result as a non-water body;
step 43: deleting the water body smaller than the area threshold value by a seed point diffusion method;
step 44: supplementing the edge part of the complete water body through a region growing algorithm;
step 45: and generating a final water body product.
The second purpose of the invention is to provide a radar data water body information extraction system based on a land surface water body product, which comprises a data acquisition module for acquiring satellite data, and further comprises the following modules:
a preprocessing module: the satellite data are preprocessed to generate backscattering data, and a derivative coefficient of the backscattering data is calculated;
a sample training module: for creating a training sample dataset using global water product data, the backscatter data and the derivative coefficients;
the information extraction module: the method is used for extracting water body information of training sample data based on a random forest model to obtain an initial water body extraction result;
a post-processing module: the water body extraction device is used for carrying out post-treatment on the initial water body extraction result to obtain a final water body product;
each module in the system performs radar data water body information extraction according to the method as described in the first aim.
Preferably, the preprocessing is to perform orbit correction, radiation correction, speckle filtering, multi-vision, terrain correction and db conversion operations on the intensity of the satellite data after the satellite data is subjected to basic processing and then to generate the backscattering data finallyVHAndVV。
in any of the above aspects, preferably, the derivative coefficient includes a polarization ratioVHrVVThe calculation formula is
。
Preferably in any of the above aspects, the derivation coefficients further include normalized deviation polarization indexNDPIThe calculation formula is
。
Preferably in any of the above aspects, the derivation coefficients further comprise normalized VH indicesNVHIThe calculation formula is
。
In any of the above aspects, preferably, the derived coefficients further include a normalized VV indexNVVIThe calculation formula is
。
In any of the above schemes, preferably, the sample training module is configured to separately train the water body sample and the non-water body sample according to theVVSorting the values, and selecting the quantiles in the water body sample asN1% as the maximum threshold, deleting samples greater than the maximum threshold; the water body sample has the following componentsN2% value as the minimum threshold, samples less than the minimum threshold are deleted.
In any of the above schemes, preferably, the information extraction includes the following sub-steps:
step 31: randomly selecting from training data setsN3% water samples, when the number of the water samples is less than a number threshold, selecting all the water samples;
step 32: determining the optimal parameters of a random forest classifier by using a cross validation grid GridSearch, and generating a training model;
step 33: applying the training model to all pixels of the whole image to obtain the probability that each pixel is a water body;
step 34: and obtaining water body distribution according to the threshold value to obtain an initial water body extraction result.
In any of the above aspects, preferably, the post-processing comprises the following sub-steps:
step 41: calculating a gradient map by using a digital elevation model DEM of srtm 30;
step 42: assigning the corresponding extraction result with the gradient larger than the gradient threshold value in the initial water body extraction result as a non-water body;
step 43: deleting the water body smaller than the area threshold value by a seed point diffusion method;
step 44: supplementing the edge part of the complete water body through a region growing algorithm;
step 45: and generating a final water body product.
The invention provides a radar data water body information extraction method and system based on a land surface water body product, which can quickly and automatically extract large-range water body information without being influenced by mountain shadow, and the extracted water bodies are relatively complete and communicated.
The seacoast data is one of JRC global water products, and records the number of times of occurrence of water in each pixel of a global land area in the last year, wherein the maximum value is 12, namely the pixel always exists in the water in the last year; the minimum is 0, namely a non-water body pixel; 1-11 are seasonal water bodies. The latest water product is seacontinity 2018 made based on the Landsat data of 2018.
Drawings
Fig. 1 is a flowchart of a preferred embodiment of a radar data water body information extraction method based on a land water body product according to the invention.
Fig. 2 is a flowchart of a preferred embodiment of the radar data water body information extraction method based on the land water body product according to the invention.
Fig. 3 is a flowchart of a preferred embodiment of the radar data water body information extraction method based on the land water body product according to the invention.
Fig. 4 is a block diagram of a preferred embodiment of the radar data water body information extraction system based on the land water body product according to the invention.
Fig. 5 is a technical route diagram of another preferred embodiment of the radar data water body information extraction method based on the land water body product according to the invention.
Fig. 6 is a cloud shadow displacement diagram of the embodiment shown in fig. 3 of the radar data water body information extraction method based on the land water body product according to the invention.
Fig. 7 is a schematic diagram of a wide river water body extraction result according to a radar data water body information extraction method based on a land water body product of the present invention. Fig. 8 is a schematic diagram of a narrow stream water body extraction result of the embodiment shown in fig. 7 of the radar data water body information extraction method based on the land water body product according to the invention.
Fig. 9 is a pre-disaster satellite data diagram of a flood event according to a preferred embodiment of the radar data water body information extraction method based on a land surface water body product of the present invention.
Fig. 10 is a diagram illustrating a pre-disaster water body extraction result according to the embodiment shown in fig. 9 of the radar data water body information extraction method based on a land water body product according to the present invention.
Fig. 11 is a diagram of the satellite data in disaster according to the embodiment shown in fig. 9 of the radar data water body information extraction method based on the land water body product according to the present invention.
Fig. 12 is a diagram illustrating a result of the extraction of water in disaster according to the embodiment shown in fig. 9 of the radar data water information extraction method based on a terrestrial water product according to the present invention.
Fig. 13 is a diagram of post-disaster satellite data according to the embodiment shown in fig. 9 of the radar data water body information extraction method based on a land water body product according to the present invention.
Fig. 14 is a diagram illustrating post-disaster water body extraction results according to the embodiment shown in fig. 9 of the radar data water body information extraction method based on a land water body product according to the present invention.
Detailed Description
The invention is further illustrated with reference to the figures and the specific examples.
Example one
As shown in fig. 1, a radar data water body information extraction method based on a terrestrial water body product executes step 100 to obtain satellite data. The satellite data refers to the data returned by the Sentinel-1A family of satellites.
Step 110 is executed to preprocess the satellite data to generate backscatter data and to calculate a coefficient of derivation of the backscatter data. The preprocessing is to carry out track correction on the strength of a grade-1 ground distance product after the satellite data are subjected to basic processing,Radiation correction, speckle filtering, multi-view, terrain correction, and db conversion operations to ultimately generate backscatter dataVHAndVV. The derivative coefficient includes polarization ratioVHrVVNormalized deviation polarization indexNDPINormalized VH indexNVHIAnd normalized VV indexNVVI. Polarization ratio ofVHrVVIs calculated by the formula
. Normalized deviation polarization indexNDPIIs calculated by the formula
Normalized VH indexNVHIIs calculated by the formula
. Normalized VV indexNVVIIs calculated by the formula
。
Step 120 is performed to create a training sample data set using the global water product data, the backscatter data, and the derivative coefficients. Respectively pressing the water body sample and the non-water body sample according to theVVSorting the values, and selecting the quantiles in the water body sample asN1% as the maximum threshold, deleting samples greater than the maximum threshold; the water body sample has the following componentsN2% value as the minimum threshold, samples less than the minimum threshold are deleted. In the present embodiment, it is preferred that,N1=85，N2=15。
and step 130, extracting water body information of the training sample data based on the random forest model to obtain an initial water body extraction result. As shown in fig. 2, step 131 is performed to randomly select training data setsN3% of water samples, when the number of water samples is less than a number threshold, all water samples are selected, in this embodiment,N3= 12.5. Step 132 is executed, the cross validation grid GridSearch is used to determine the optimal parameters of the random forest classifier, and a training model is generated. Step 133 is performed to apply the training model to all pixels of the entire imageAnd obtaining the probability that each pixel is the water body. And step 134, obtaining water body distribution according to the threshold value to obtain an initial water body extraction result.
And 140, performing post-processing on the initial water body extraction result to obtain a final water body product. As shown in fig. 3, step 141 is executed to calculate a grade map using the digital elevation model DEM of srtm 30. And 142, assigning the corresponding extraction result with the gradient larger than the gradient threshold value in the initial water body extraction result as the non-water body. Step 143 is executed to delete the water bodies smaller than the area threshold value by the method of seed point diffusion. Step 144 is performed to supplement the edge portion of the complete body of water by a region growing algorithm. Step 145 is performed to generate the final water product.
Example two
As shown in fig. 4, a radar data water body information extraction system based on a land surface water body product includes a data acquisition module 200, a preprocessing module 210, a sample training module 220, an information extraction module 230, and a post-processing module 240.
The data acquisition module 200: for acquiring satellite data.
The preprocessing module 210: and the satellite data are preprocessed to generate backscattering data, and the derivative coefficient of the backscattering data is calculated. The preprocessing is to carry out orbit correction, radiation correction, speckle filtering, multi-vision, terrain correction and conversion operations on the intensity of the satellite data after basic processing and a grade-1 ground distance product. The derivative coefficient includes polarization ratioVHrVVNormalized deviation polarization indexNDPINormalized VH indexNVHIAnd normalized VV indexNVVI. Polarization ratio ofVHrVVIs calculated by the formula
. Normalized deviation polarization indexNDPIIs calculated by the formula
. Normalized VH indexNVHIIs calculated by the formula
. Normalized VV indexNVVIIs calculated by the formula
。
The sample training module 220: for creating a training sample data set using the global water product data, the backscatter data and the derivative coefficients. Respectively pressing the water body sample and the non-water body sample according to theVVSorting the values, and selecting the quantiles in the water body sample asN1% as the maximum threshold, deleting samples greater than the maximum threshold; the water body sample has the following componentsN2% value as the minimum threshold, samples less than the minimum threshold are deleted.
The information extraction module 230: and the method is used for extracting the water body information of the training sample data based on the random forest model to obtain an initial water body extraction result. The information extraction comprises the following substeps: step 31: randomly selecting from training data setsN3% water samples, when the number of the water samples is less than a number threshold, selecting all the water samples; step 32: determining the optimal parameters of a random forest classifier by using a cross validation grid GridSearch, and generating a training model; step 33: applying the training model to all pixels of the whole image to obtain the probability that each pixel is a water body; step 34: and obtaining water body distribution according to the threshold value to obtain an initial water body extraction result.
The post-processing module 240: and the water body extraction device is used for carrying out post-treatment on the initial water body extraction result to obtain a final water body product. The post-processing comprises the following sub-steps: step 41: calculating a gradient map by using a digital elevation model DEM of srtm 30; step 42: assigning the corresponding extraction result with the gradient larger than the gradient threshold value in the initial water body extraction result as a non-water body; step 43: deleting the water body smaller than the area threshold value by a seed point diffusion method; step 44: supplementing the edge part of the complete water body through a region growing algorithm; step 45: and generating a final water body product.
EXAMPLE III
The automatic extraction of the water body information provided by the invention mainly comprises four steps:
(1) preprocessing the Sentinel-1 SAR data to generate backscattering data and calculating a derivative coefficient of the backscattering data;
(2) creating a training sample data set by using the seasconity 2018 product, the backscattering data and the derivative coefficient;
(3) extracting water body information based on a random forest model;
(4) and (3) post-treatment: firstly removing the water body with obvious misclassification, then removing mountain shadow by using gradient data, then removing small communicated regions, and supplementing the water body by using region growth to obtain a final water body product; and finally, carrying out precision evaluation on the product by using the auxiliary image.
The technical route is shown in fig. 5.
1. Sar data preprocessing and index calculation
The intensity of a High-resolution group range Detected (5 mx20 m) Level-1 ground detection product GRDH of the Sentinal-1 is subjected to track correction, radiation correction, speckle filtering, multi-vision and terrain correction, and is converted into dB. The speckle filtering adopts Refine-Lee filtering, the effect of converting an image into 20m square pixels in multiple views is good (more speckle noise is generated at 10m, and lower resolution is generated at 30 m), and a digital elevation model of SRTM 1 arc second (about 30 m) is used for terrain correction.
Correlation index calculations are performed simultaneously and added to the classification as feature data. The polarization ratio VHrVV, the normalized deviation polarization index NDPI, the normalized VH index NVHI and the normalized VV index NVVI are included. The index calculation is shown in table 1.
TABLE 1 list of polarization indexes
2. Water body training sample creation
The sample data is derived from seakeeping data of JRC Global Surface Water. The data records the number of months of occurrence of each water body pixel in the inland of the global region in the last year, and the maximum value is 12, namely the water bodies exist all the time in the year; the lowest is 0, namely the non-water pixel; 1-11 are seasonal water bodies. The latest product is seaquality 2018. Due to the problems of seasons and precision, fewer misclassified samples can appear in the water body samples and the non-water body samples, so that the water body samples and the non-water body samples are firstly sorted according to the VV value by combining the characteristics of the water body in a radar image, then the value of 85% of the quantiles in the water body samples is selected as the maximum threshold value, and the samples larger than the maximum threshold value are deleted; and taking the value with 15% of the quantiles in the non-water body sample as a minimum threshold value, and deleting the samples smaller than the minimum threshold value.
3. Regional water body information extraction
A random forest model was used to classify bodies of water from the Sentinel-1 data. Random forests have excellent accuracy in all current algorithms. The classification method of machine learning is influenced by unbalanced training data, and the prediction effect is not good for classes with few samples. So the ratio of water to non-water is greater than 1: 20, herein limited to 1: 20. in order to simplify the training samples, 1/8 water body samples in the training data set are randomly sampled, when the water body samples are less than 10000, all the water body samples are selected, and the non-water body samples are randomly selected according to the proportion. The cross-validation grid is also used to determine the best parameters for the random forest classifier, such as the number of trees and the maximum number of features of the trees used. And then applying the trained model to all pixels of the whole image to obtain the probability that each pixel is the water body. And finally, obtaining water body distribution according to a threshold, wherein the water body is selected to be more than 0.5, so that an initial water body extraction result is obtained.
4. Post-treatment
After information extraction, a small number of abnormal high values are wrongly classified into water bodies, and in order to remove the influence of the part of values, the water bodies with VV values larger than the maximum threshold value of the water bodies are assigned as non-water bodies. After data after random forests are classified, most of water bodies are extracted, however, due to the characteristics of radar images, mountain shadows also present the same characteristics as the water bodies, and for all water bodies, when the corresponding gradient of the gradient map is larger than 5 degrees, the gradient map is assigned as a non-water body, namely larger than a certain gradient, the water bodies cannot be retained. The water product generated here has more scattered pixel points, and we delete the water smaller than a certain area by a seed point diffusion method. Due to the problems of the DEM precision and errors, the edge part of the complete water body is removed, and the water body is supplemented through a region growing algorithm, so that a final water body product is obtained.
Example four
In the embodiment, the coastal flood events of Taizhou city of Zhejiang province of 8 months and 10 days in 2019 are selected as research objects, the Zhejiang coastal flood comprises 3 counties, the average elevation in the area is 300m, the terrain is complex, the mountain area and the plain area exist, the water body types in the area range are various, and the water collection area is about 254 square kilometers.
The data used are Sentinil-1 SAR, Landsat8, JRCGlobal Surface Water, SRTM DEM data.
The Sentinel-1 satellite is the first copernik planned satellite constellation transmitted by the european space, and is combined by two satellites A, B, so that the image acquisition of the same place can be realized every 6 days under the combination of the two satellites. It is a C-band satellite, with four imaging modes: hyperfine Mode (SM), Interferometric Wide Mode (IW), ultra-Wide Mode (EW), microwave Mode (Wave-Mode). The main modes in which earth and land are covered are the IW mode, the SM mode is mainly used for emergency events, and the EW and Wave are mainly used for marine monitoring. The Sentinel-1 satellite data can be freely acquired, the width of an IW mode is 250km, the resolution is 5x20m, the data are shot from a satellite and distributed to a database for only about 3-6 hours, cloud and fog can be penetrated through in-ground observation, weather images are not received, and the characteristics enable the Setinel-1 satellite data to be very suitable for flood disaster remote sensing monitoring. The Sentinel-1 data herein is downloaded from the european space website (https:// scihub. copernius. eu /), and also from NASA-forwarded websites (https:// search. asf. alaska. edu).
The JRC global surface water product is mainly used for manufacturing training samples for calibrating random forest models. The aquatic product was produced by the european union research center and contained a map of the location and time of surface water from 1984 to 2018 and provided range and statistical data of water surface generated from 3,865,618 scenes collected from Landsat 5, 7 and 8 using 16 days 3 and 2018 and 31 days 12 and 8 in 1984. The expert system is used to classify each pixel as water/non-water, respectively, and collate the results into a monthly history for the change detection over the entire period and for two periods (1984-1999, 2000-2018). The data used primarily herein is seaselectivity 2018, which records primarily the appearance of water for months in 2018, with 12 months being permanent water and less than 12 months being seasonal water.
EXAMPLE five
As shown in fig. 6, the water body extraction result of the fourth example is shown, most water body areas are extracted, and the extraction missing phenomenon is less and is not influenced by mountain shadow basically. In fig. 7, the wide river water body in the local area 1 is extracted relatively completely and is not affected by the mountain shadow at the upper right corner; the narrower streams and ponds in local area 2 of fig. 8 are also extracted and the river is more complete.
EXAMPLE six
As shown in fig. 9-14, the satellite images and the water body extraction results before, during and after the flood in the near-sea flood event in Zhejiang are respectively shown, and the results show that the water body extraction is relatively complete, and can be better applied to flood disaster remote sensing monitoring.
For a better understanding of the present invention, the foregoing detailed description has been given in conjunction with specific embodiments thereof, but not with the intention of limiting the invention thereto. Any simple modifications of the above embodiments according to the technical essence of the present invention still fall within the scope of the technical solution of the present invention. In the present specification, each embodiment is described with emphasis on differences from other embodiments, and the same or similar parts between the respective embodiments may be referred to each other. For the system embodiment, since it basically corresponds to the method embodiment, the description is relatively simple, and for the relevant points, reference may be made to the partial description of the method embodiment.
Claims (10)
1. A radar data water body information extraction method based on a land surface water body product comprises the steps of obtaining satellite data, and is characterized by further comprising the following steps:
step 1: preprocessing the satellite data to generate backscattering data and calculating a derivative coefficient of the backscattering data;
step 2: creating a training sample dataset using global water product data, the backscatter data and the derivative coefficients;
and step 3: extracting water body information of the training sample data based on a random forest model to obtain an initial water body extraction result;
and 4, step 4: and carrying out post-treatment on the initial water body extraction result to obtain a final water body product.
2. The method as claimed in claim 1, wherein the preprocessing comprises performing orbit correction, radiation correction, speckle filtering, multi-vision, terrain correction and db conversion on the intensity of the level 1 product after the satellite data is subjected to basic processing, and finally generating the back scattering dataVHAndVV。
3. the method of claim 2, wherein the derivative coefficients comprise polarization ratiosVHrVVThe calculation formula is
。
4. The method of claim 2, wherein the derivative coefficients further comprise normalized deviation polarization indexNDPIThe calculation formula is
。
5. The method of claim 2, wherein the derivation coefficients further comprise normalized VH indicesNVHIThe calculation formula is
。
6. The method of claim 2, wherein the derivation coefficients further comprise a normalized VV indexNVVIThe calculation formula is
。
7. The method of claim 2, wherein the step 2 comprises selecting the water body sample and the non-water body sample according to the water body informationVVSorting the values, and selecting the quantiles in the water body sample asN1% as the maximum threshold, deleting samples greater than the maximum threshold; the water body sample has the following componentsN2% value as the minimum threshold, samples less than the minimum threshold are deleted.
8. The radar-data water body information extraction method based on land-based water body products according to claim 7, wherein the step 3 comprises the following sub-steps:
step 31: randomly selecting from training data setsN3% water samples, when the number of the water samples is less than a number threshold, selecting all the water samples;
step 32: determining the optimal parameters of a random forest classifier by using a cross validation grid GridSearch, and generating a training model;
step 33: applying the training model to all pixels of the whole image to obtain the probability that each pixel is a water body;
step 34: and obtaining water body distribution according to the threshold value to obtain an initial water body extraction result.
9. The radar-data water body information extraction method based on land-based water body products according to claim 8, wherein the step 4 comprises the following sub-steps:
step 41: calculating a gradient map by using a digital elevation model DEM of srtm 30;
step 42: assigning the corresponding extraction result with the gradient larger than the gradient threshold value in the initial water body extraction result as a non-water body;
step 43: deleting the water body smaller than the area threshold value by a seed point diffusion method;
step 44: supplementing the edge part of the complete water body through a region growing algorithm;
step 45: and generating a final water body product.
10. A radar data water body information extraction system based on a land surface water body product comprises a data acquisition module for acquiring satellite data, and is characterized by further comprising the following modules:
a preprocessing module: the satellite data are preprocessed to generate backscattering data, and a derivative coefficient of the backscattering data is calculated;
a sample training module: for creating a training sample dataset using global water product data, the backscatter data and the derivative coefficients;
the information extraction module: the method is used for extracting water body information of training sample data based on a random forest model to obtain an initial water body extraction result;
a post-processing module: the water body extraction device is used for carrying out post-treatment on the initial water body extraction result to obtain a final water body product;
modules in the system perform radar data water body information extraction according to the method of claim 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010370050.4A CN111259876B (en) | 2020-05-06 | 2020-05-06 | Radar data water body information extraction method and system based on land surface water body product |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010370050.4A CN111259876B (en) | 2020-05-06 | 2020-05-06 | Radar data water body information extraction method and system based on land surface water body product |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111259876A true CN111259876A (en) | 2020-06-09 |
| CN111259876B CN111259876B (en) | 2020-10-09 |
Family
ID=70951695
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010370050.4A Active CN111259876B (en) | 2020-05-06 | 2020-05-06 | Radar data water body information extraction method and system based on land surface water body product |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111259876B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111652882A (en) * | 2020-07-07 | 2020-09-11 | 中国水利水电科学研究院 | Large-scale surface water product drawing precision evaluation method |
| CN111738144A (en) * | 2020-06-19 | 2020-10-02 | 中国水利水电科学研究院 | Surface water product generation method and system based on Google Earth Engine cloud platform |
| CN111931709A (en) * | 2020-09-17 | 2020-11-13 | 航天宏图信息技术股份有限公司 | Water body extraction method and device for remote sensing image, electronic equipment and storage medium |
| CN113111567A (en) * | 2021-02-26 | 2021-07-13 | 中国科学院空天信息创新研究院 | Evapotranspiration component separation method, component separation device, apparatus, and medium |
| CN113191292A (en) * | 2021-05-11 | 2021-07-30 | 四川大学 | Method for rapidly extracting bursting flood submerging range |
| CN113505764A (en) * | 2021-09-09 | 2021-10-15 | 成都理工大学 | Local incidence angle considered method for identifying SAR geometric distortion of high mountain canyon region |
| CN113567981A (en) * | 2021-06-28 | 2021-10-29 | 中国电建集团华东勘测设计研究院有限公司 | SAR image-based flood risk area automatic extraction method |
| CN113642399A (en) * | 2021-07-12 | 2021-11-12 | 广东省国土资源测绘院 | Method for identifying abandoned land in paddy field based on SAR data |
| CN114677401A (en) * | 2022-03-28 | 2022-06-28 | 珠江水利委员会珠江水利科学研究院 | Water body extraction method and system based on polarization radar self-image features |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109472804A (en) * | 2018-11-05 | 2019-03-15 | 南方科技大学 | Land table Clean water withdraw method and apparatus based on remote sensing image |
-
2020
- 2020-05-06 CN CN202010370050.4A patent/CN111259876B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109472804A (en) * | 2018-11-05 | 2019-03-15 | 南方科技大学 | Land table Clean water withdraw method and apparatus based on remote sensing image |
Non-Patent Citations (2)
| Title |
|---|
| WENLI HUANG等: "Automated Extraction of Surface Water Extent from Sentinel-1 Data", 《REMOTE SENSING》 * |
| 孙亚勇等: "基于星载SAR数据的山区水体提取方法研究", 《中国水利水电科学研究院学报》 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111738144A (en) * | 2020-06-19 | 2020-10-02 | 中国水利水电科学研究院 | Surface water product generation method and system based on Google Earth Engine cloud platform |
| CN111738144B (en) * | 2020-06-19 | 2021-08-31 | 中国水利水电科学研究院 | Surface water product generation method and system based on Google Earth Engine cloud platform |
| CN111652882A (en) * | 2020-07-07 | 2020-09-11 | 中国水利水电科学研究院 | Large-scale surface water product drawing precision evaluation method |
| CN111931709B (en) * | 2020-09-17 | 2021-01-05 | 航天宏图信息技术股份有限公司 | Water body extraction method and device for remote sensing image, electronic equipment and storage medium |
| CN111931709A (en) * | 2020-09-17 | 2020-11-13 | 航天宏图信息技术股份有限公司 | Water body extraction method and device for remote sensing image, electronic equipment and storage medium |
| CN113111567A (en) * | 2021-02-26 | 2021-07-13 | 中国科学院空天信息创新研究院 | Evapotranspiration component separation method, component separation device, apparatus, and medium |
| CN113191292A (en) * | 2021-05-11 | 2021-07-30 | 四川大学 | Method for rapidly extracting bursting flood submerging range |
| CN113191292B (en) * | 2021-05-11 | 2022-01-04 | 四川大学 | Method for rapidly extracting bursting flood submerging range |
| CN113567981A (en) * | 2021-06-28 | 2021-10-29 | 中国电建集团华东勘测设计研究院有限公司 | SAR image-based flood risk area automatic extraction method |
| CN113567981B (en) * | 2021-06-28 | 2023-08-08 | 中国电建集团华东勘测设计研究院有限公司 | Automatic extraction method for flood risk area based on SAR image |
| CN113642399A (en) * | 2021-07-12 | 2021-11-12 | 广东省国土资源测绘院 | Method for identifying abandoned land in paddy field based on SAR data |
| CN113505764A (en) * | 2021-09-09 | 2021-10-15 | 成都理工大学 | Local incidence angle considered method for identifying SAR geometric distortion of high mountain canyon region |
| CN114677401A (en) * | 2022-03-28 | 2022-06-28 | 珠江水利委员会珠江水利科学研究院 | Water body extraction method and system based on polarization radar self-image features |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111259876B (en) | 2020-10-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111259876B (en) | Radar data water body information extraction method and system based on land surface water body product | |
| Saha et al. | Land cover classification using IRS LISS III image and DEM in a rugged terrain: a case study in Himalayas | |
| CN109919875B (en) | High-time-frequency remote sensing image feature-assisted residential area extraction and classification method | |
| Borak | Feature selection and land cover classification of a MODIS-like data set for a semiarid environment | |
| Sowmya et al. | Remote sensing satellite image processing techniques for image classification: a comprehensive survey | |
| CN107507200B (en) | SAR image high-precision large-range water area extraction method based on connectivity detection and noise suppression | |
| Zhang et al. | GWL_FCS30: global 30 m wetland map with fine classification system using multi-sourced and time-series remote sensing imagery in 2020 | |
| US20230186503A1 (en) | Method for troubleshooting hidden dangers of trees near power transmission lines by combining icesat-2 with jl-1 images | |
| Saeed et al. | Coastline extraction using satellite imagery and image processing techniques | |
| Liao et al. | Urban change detection based on coherence and intensity characteristics of SAR imagery | |
| CN112419197A (en) | Universal single-time phase and multi-time phase SAR image speckle noise removing method | |
| Wang et al. | Arctic sea ice cover data from spaceborne SAR by deep learning | |
| Guo et al. | Mozambique flood (2019) caused by tropical cyclone idai monitored from sentinel-1 and sentinel-2 images | |
| Tavus et al. | A Cnn-based flood mapping approach using sentinel-1 data | |
| Kumar et al. | Multi-sensor multi-resolution image fusion for improved vegetation and urban area classification | |
| Li et al. | Evaluating the ability of NOAA-20 monthly composite data for socioeconomic indicators estimation and urban area extraction | |
| Wu et al. | ESRGAN-based DEM super-resolution for enhanced slope deformation monitoring in lantau island of Hong Kong | |
| Haris et al. | COMPARISON OF SENTINEL-2 AND MULTITEMPORAL SENTINEL-1 SAR IMAGERY FOR MAPPING AQUACULTURE POND DISTRIBUTION IN THE COASTAL REGION OF BREBES REGENCY, CENTRAL JAVA, INDONESIA. | |
| Borde et al. | A direct link between feature tracking and height assignment of operational atmospheric motion vectors | |
| Li et al. | “3S” TECHNOLOGIES AND APPLICATION FOR DYNAMIC MONITORING SOIL AND WATER LOSS IN THE YANGTZE RIVER BASIN, CHINA | |
| Wang et al. | Framework to create cloud-free remote sensing data using passenger aircraft as the platform | |
| Sasaki et al. | Enhancing the Detection of Coastal Marine Debris in Very High Resolution Satellite Imagery via Unsupervised Domain Adaptation | |
| Plutalova et al. | Central Yamal vegetation monitoring based on Sentinel-2 and Sentinel-1 imagery | |
| Mladinich | An evaluation of object-oriented image analysis techniques to identify motorized vehicle effects in semi-arid to arid ecosystems of the American West | |
| Hong et al. | Automatic mapping technique of sea ice in the coastal waters of the Arctic Ocean using Sentinel-1 data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |