CN113567981A - SAR image-based flood risk area automatic extraction method - Google Patents

SAR image-based flood risk area automatic extraction method Download PDF

Info

Publication number
CN113567981A
CN113567981A CN202110718205.3A CN202110718205A CN113567981A CN 113567981 A CN113567981 A CN 113567981A CN 202110718205 A CN202110718205 A CN 202110718205A CN 113567981 A CN113567981 A CN 113567981A
Authority
CN
China
Prior art keywords
water body
sigma
probability
flood
risk area
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
Application number
CN202110718205.3A
Other languages
Chinese (zh)
Other versions
CN113567981B (en
Inventor
白珏莹
邬雪松
郭靖
富强
张善亮
房新力
刘强
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
PowerChina Huadong Engineering Corp Ltd
Original Assignee
PowerChina Huadong Engineering Corp Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by PowerChina Huadong Engineering Corp Ltd filed Critical PowerChina Huadong Engineering Corp Ltd
Priority to CN202110718205.3A priority Critical patent/CN113567981B/en
Publication of CN113567981A publication Critical patent/CN113567981A/en
Application granted granted Critical
Publication of CN113567981B publication Critical patent/CN113567981B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
    • G01S13/88—Radar or analogous systems specially adapted for specific applications
    • G01S13/89—Radar or analogous systems specially adapted for specific applications for mapping or imaging
    • G01S13/90—Radar or analogous systems specially adapted for specific applications for mapping or imaging using synthetic aperture techniques, e.g. synthetic aperture radar [SAR] techniques
    • G01S13/9021—SAR image post-processing techniques
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
    • Y02A10/40—Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping

Landscapes

  • Engineering & Computer Science (AREA)
  • Remote Sensing (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Physics & Mathematics (AREA)
  • Electromagnetism (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • General Physics & Mathematics (AREA)
  • Image Analysis (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

The invention discloses an automatic flood risk area extraction method based on SAR images, which utilizes a C-band backscattering coefficient sigma0And calculating the water body probability of the SAR images before and after the flood event in two time phases, and then expressing the flood risk area by using the water body probability increment. The method is a rapid, convenient and near-real-time flood monitoring method, and can effectively extract newly added flooding areas (flood risk areas) before and after a rainfall event. The method comprises the following steps: acquiring two time phase images before and after the flood event to form sigma0Gathering; generating sigma0And assuming sigma0Conforming to the Gaussian mixture distribution, and then filtering the outlier; for down-sampled sigma0Carrying out unsupervised classification on the set, and calculating prior probability; down-sampled sigma using maximum expected EM algorithm and prior probability0Performing iterative calculation on the set, calculating the parameters of the distribution curve, and based on sigma before down-sampling0Calculating the probability of the water body; and extracting the flood risk area by using the water probability increment of the two time phase images.

Description

SAR image-based flood risk area automatic extraction method
Technical Field
The invention belongs to the technical field of flood monitoring, and particularly discloses an automatic extraction method of a flood risk area based on an SAR image. The method utilizes the C-band backscattering coefficient (sigma)0) Calculating the water body probability of the SAR images in two time phases before and after the flood event, and then expressing the newly added flood inundation area (flood risk area) after the rainfall event by using the water body probability increment of the images in the two time phases. The method is a simple, rapid and effective flood monitoring method, can rapidly monitor the area change of the water body, and can automatically extract the flood risk area.
Background
Global climate warming, frequent occurrence of abnormal climate phenomena, gradual frequent flood disasters and great influence on the life health and living environment of human beings. Particularly, in southern areas of China, the areas are mostly in tropical and subtropical monsoon climates, and severe flood disasters are caused to the areas by strong rainfall and upstream ice and snow ablation phenomena. Therefore, the method is very important for effectively and quickly detecting the water body change in the flood season and automatically extracting the flood risk area.
Under heavy rainfall and overcast and rainy weather, the imaging conditions of the optical images are poor, a large amount of data is lost, and the application of the optical images in the detection of the water body change in the flood season is restricted. The Synthetic Aperture Radar (SAR) is not influenced by overcast and rainy, has all-weather earth observation capability all day long, and has great potential in water body range extraction and flood risk area extraction. Currently, there are a large number of SAR satellites operating in orbit, such as: the Sentinel-1 satellite, the GF-3 satellite and the like are excellent SAR image data sources, and provide important data support for flood monitoring in China.
In general, the flood risk area extraction based on remote sensing data is to extract a water body part of two time-phase images, and then perform superposition analysis to analyze water body changes before and after a flood. Under the condition of very complex water environment, the phenomena of wrong extraction, missed detection and the like are easy to occur in water extraction, and the uncertainty of the type of the ground object to which the pixel belongs is one of the reasons to a certain extent. In addition, research is also available, in which operations are performed on image backscattering at different periods, and then water body change detection is performed on the operation results by using methods such as threshold segmentation and the like, so that post-disaster evaluation and flood risk area extraction are performed. Due to the complexity of the water environment, the threshold segmentation may also have some uncertainty, and the threshold setting is also lack of automation.
Disclosure of Invention
The invention aims to provide a flood risk area extraction method, which is used for directly monitoring flood based on two-time-phase SAR images before and after a flood event, and constructing the flood risk area extraction method based on water probability by calculating the water probability increment after the flood event, so that the flood can be quickly monitored.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
an automatic flood risk area extraction method based on SAR images is characterized by comprising the following steps:
step (1), acquiring two time phase images before and after an flood event in an area, preprocessing the two time phase images, and forming a backscattering coefficient (sigma) by all pixels0) Set of sigma1；
Step (2) for set Σ1Drawing a frequency histogram, and assuming that sigma of two-time phase SAR images distributed with water body0The set conforms to the Gaussian mixture distribution and is aligned with sigma0Filtering outliers in the set;
step (3) of resampling the two time-phase SAR images filtered in the step (2) with reduced spatial resolution to generate sigma0Set of sigma2And assuming downsampled sigma2In a mixed gaussian distribution, and the resulting distribution can be used to downsample the pre-set sigma1The probability of the water body is calculated, and the sigma after the down sampling is calculated0Set of sigma2Performing k-means clustering, and calculating the prior probability of the water body and the non-water body;
step (4), utilizing the maximum Expectation (EM) algorithm and the water body prior probability in the step (3) to carry out the down-sampling on the set sigma2Performing iterative calculation on the distribution parameters to increase the iteration speedDegree, then σ for each pixel before down-sampling0Calculating the water probability according to the value;
and (5) expressing the change value of the probability of each pixel belonging to the water body before and after the event by using the water body probability increment, determining the water body probability increment boundary value of the flood risk area and the non-risk area, and constructing a flood risk area extraction method based on the two-time phase water body probability.
On the basis of the technical scheme, the invention can also adopt the following further technical schemes or combine the further technical schemes for use:
the step (1) specifically comprises the following substeps:
step (11), acquiring SAR images covering a research area before and after a flooding event;
step (12), registration, multi-temporal filtering, geocoding, radiometric calibration and spatial cutting are carried out on the two-temporal SAR images by using Sarscape software to obtain sigma0(unit: dB);
step (13), all pixels of the two-time phase image are combined to form sigma0Set of sigma1。
In the step (2), the sets Σ1Drawing a frequency histogram, and assuming that sigma of two-time phase SAR images distributed with water body0The set conforms to the Gaussian mixture distribution, and the outlier is filtered, specifically comprising the following substeps:
step (21) for σ0Drawing a frequency histogram in a set manner, and assuming that a probability density curve of the frequency histogram accords with mixed Gaussian distribution;
and (22) even if a series of preprocessing is carried out on the SAR image by using the terrain data and the two-time phase characteristics, a larger abnormal value still exists to influence subsequent parameter estimation and water body probability calculation, and therefore, outliers are removed by arranging the values from small to large.
The step (3) specifically comprises the following substeps:
step (31), in order to increase the iterative computation speed in step (4), the two-time phase SAR image filtered in step (2) is down-sampled to generate sigma0Set of sigma2And assuming a setΣ2The probability density function of (a) also conforms to a Gaussian mixture distribution and is based on set sigma2The probability density function of (2) may be applied to the sigma of step (1)0And calculating the probability of the water body.
Step (32), for sigma0Set of sigma2K mean value clustering is carried out, and sigma of water body class is collected0Sigma of aggregate and non-water classes0Gathering;
step (33), calculating the prior probability of the water body and the non-water body; the ratio of the number of the pixels in the water body pixel set to the number of all the pixels is the prior probability p of the water body0(W) and p0(W) is an initial value of p (W), the ratio of the number of the pixels of the non-water body pixel set to the number of all the pixels is a non-water body prior probability, and the formula is as follows:
Figure BDA0003135844920000031
wherein W is a water body pixel set,
Figure BDA0003135844920000032
is a non-aqueous body pixel set, p (W) and
Figure BDA0003135844920000033
respectively representing the water body prior probability and the non-water body prior probability, and the sum of the two is equal to 1.
The step (4) specifically comprises the following substeps:
step (41), constructing sigma by using mixed Gaussian distribution0A probability density function of the marginal distribution, the density function being as follows:
Figure BDA0003135844920000034
wherein W is a water body pixel set,
Figure BDA0003135844920000041
is a non-aqueous body pixel set, p (W) and
Figure BDA0003135844920000042
respectively representing the prior probability of water and the prior probability of non-water, p (sigma)0) Representing the probability that the pixel corresponding to the backscattering coefficient belongs to the water body;
Figure BDA0003135844920000043
Figure BDA0003135844920000044
wherein, muW、sW、
Figure BDA0003135844920000045
Are respectively water body pixel sigma0Mean value, standard deviation and non-water body pixel sigma0Mean, standard deviation of;
and (42) iteratively calculating the mean value and standard deviation parameters of the distribution curve by using a maximum Expectation (EM) algorithm based on maximum likelihood estimation, and iteratively optimizing p (W), wherein the formula is as follows:
Figure BDA0003135844920000046
wherein EM (DEG) represents the maximum Expectation (EM) algorithm, p (W) is the water body prior probability obtained by k-means clustering, sigma0As backscatter coefficient data,. mu.W、sW、
Figure BDA0003135844920000047
Are respectively water body pixel sigma0Mean value, standard deviation and non-water body pixel sigma0Mean, standard deviation of.
A step (43) of converting p (W) optimized in the step (42),
Figure BDA0003135844920000048
p(σ0I W) and
Figure BDA0003135844920000049
substituting the water body probability calculation formula to construct a water body probability calculation formula, wherein the formula is as follows:
p(W|σ0)＝p(W)p(σ0|W)/p(σ0) (7)
step (44) of down-sampling the sigma of the two-phase image before down-sampling0Substituting into a formula, and respectively calculating the water body probability of the two time phase images.
The step (5) specifically comprises the following substeps:
step (51), the difference of the water body probabilities before and after the flood event is obtained, wherein the water body probability before the flood event is recorded as p1And the probability of the water body after the flood event is recorded as p2Subtracting the water body probability value before the flood event from the water body probability value after the flood event, and recording as p21；
In step (52), the backscattering is rough due to wind and the like, so that the backscattering coefficient is increased, but the water body before the rainfall event is always kept as the water body after the rainfall event under the normal condition, and p is ideally kept21There are no negative values, so 0 is taken to be p21The method comprises the following steps of (1) extracting a flood risk area by using water body probability, and expressing the flood risk area in a probability increment mode, wherein the extraction rule is as follows:
Figure BDA0003135844920000051
the area corresponding to the Y being larger than 0 is the flood risk area, the closer the Y is to 1, the more obvious the change of the water body submerging condition is, and the area with the Y being equal to 0 is the non-flood risk area.
And (53) further performing threshold segmentation on the Y, setting 0.2, 0.4, 0.6 and 0.8 as thresholds, wherein 0-0.2 corresponds to a first level of risk, 0.2-0.4 corresponds to a second level of risk, 0.4-0.6 corresponds to a third level of risk, 0.6-0.8 corresponds to a fourth level of risk, and 0.8-1.0 corresponds to a fifth level of risk.
The method is used for extracting the flood risk area directly based on the SAR image and utilizing the C-band backscattering coefficient (sigma)0) Computing floodingAnd (3) expressing the flood risk area by using the water body probability increment of the two time-phase SAR images before and after the event. Wherein, the probability of the water body is the probability that the pixel belongs to the water body. The method is a simple and rapid flood monitoring method, and can rapidly and automatically extract a newly increased flood inundation area (flood risk area) after a rainfall event. Compared with the prior art, the invention has the following characteristics and beneficial effects:
1. the method has the characteristics of automatically, quickly and efficiently extracting the flood risk area, does not need to manually select and set a threshold value, utilizes the mixed Gaussian distribution characteristic of two-time-phase backscattering coefficients, realizes the efficiency improvement based on the iterative calculation of downsampling data, constructs a water body probability automatic calculation method based on unsupervised classification and EM algorithm, and formulates a flood risk area extraction rule based on the water body probability to realize the automatic extraction of the newly added flood risk area.
2. The method uses the probability increment of the water body to express the change value of the probability that each pixel belongs to the water body before and after the event, compared with the change of the prior binaryzation submerging area, the method can express the uncertainty of the submerging condition of the water body on the land surface and the uncertainty of the submerging degree caused by the complex water body environment better, and in addition, the visualization effect of the flood risk area is stronger and the content is richer by expressing in the form of the probability increment.
Drawings
Fig. 1 is a calculation flow chart of an automatic flood risk area extraction method based on an SAR image according to the present invention.
Fig. 2 is a frequency histogram of a two-phase set of backscatter coefficients computed by the present invention.
Fig. 3 is water body probability values respectively corresponding to two time-phase SAR images processed by the present invention, wherein the left side and the right side respectively show water body probability values of 26 days in 6/2020 and 7/8 months in 2020.
Fig. 4 is a spatial distribution diagram of a flood risk area calculated based on water body probabilities at two time phases before and after a rainfall event according to the present invention.
Detailed Description
The invention is further explained below with reference to the drawings and the embodiments. It should be noted that the embodiments described herein are only for illustrating the present invention and are not to be construed as limiting the present invention.
The method utilizes the Sentinel-1SAR data to extract the flood risk area, the observation time of the two time phase images is 26 days 6 and 8 days 7 in 2020 respectively, and the method comprises the following steps:
step (1), acquiring two-time-phase Sentinel-1SAR images before and after a rainfall event in a local area of a Yangtze lake, preprocessing the images to obtain a 10 m-10 m spatial resolution image map, and forming a backscattering coefficient (sigma) by all pixels in a research area0) Set of sigma10*10；
Step (2) for set Σ10*10Drawing a frequency histogram, and assuming that sigma of two-time phase SAR images distributed with water body0The set conforms to the Gaussian mixture distribution and is aligned with sigma0Filtering outliers in the set;
and (3) performing 5 × 5 down-sampling on the corresponding two-time phase SAR image filtered in the step (2) to 50m × 50m to generate sigma0Set of sigma50*50To σ0Set of sigma50*50K mean value clustering is carried out, and the prior probability of the water body and the non-water body of the experimental data is calculated;
step (4), utilizing the maximum Expectation (EM) algorithm and the water body prior probability in the step (3) to carry out down-sampling on the set sigma50*50Performing iterative computation distribution parameters to improve the iterative speed, and then performing water probability computation on each pixel of 10m × 10 m;
and (5) expressing the change value of the probability of each pixel belonging to the water body before and after the event by using the water body probability increment, determining the water body probability increment boundary value of the flood risk area and the non-risk area, and constructing a flood risk area extraction method based on the two-time phase water body probability.
The invention is described in further detail below with reference to fig. 1:
in the step (1), two time phase images before and after the regional flood event are obtained, namely 26 days 6 and 7 days 8 days 7 in 2020, the images with the spatial resolution of 10m x 10m are obtained after pretreatment, and all pixels in a research region form a backscattering coefficient (sigma)0) Set of sigma10*10The method specifically comprises the following substeps:
step (11), acquiring ground distance (GRD) products of two time phases Sentinel-1 Interference Wide (IW) covering a research area before and after a flood event;
step (12), registration, multi-temporal filtering, geocoding, radiometric calibration and spatial cutting are carried out on the two-temporal SAR images by utilizing Sarscape software, and vertical-vertical (VV) polarization sigma in the research region is obtained0Image (unit: dB), here, the spatial reference is WGS84, spatial resolution 10 m;
step (13), all pixels of the two-time phase image are combined into sigma0Set of sigma10*10。
In step (2), for the set Σ10*10Drawing a frequency histogram, and assuming that sigma of two-time phase SAR images distributed with water body0The set conforms to the Gaussian mixture distribution, and the outlier is filtered, specifically comprising the following substeps:
step (21) for σ0Drawing a frequency histogram in a set manner, observing the distribution of the frequency histogram, and assuming that a probability density curve of the frequency histogram accords with mixed Gaussian distribution;
step (22), arranging the numerical values from small to large, regarding the last 0.1% of samples as outliers, filtering the outliers, and forming a mixed Gaussian distribution obtained by fusing two Gaussian distributions by using a frequency histogram after filtering the outliers as shown in FIG. 2;
the step (3) specifically comprises the following substeps:
step (31), the two time phase SAR images corresponding to the outlier samples filtered in the step (2), namely the two time phase SAR images left after 0.1% of the outlier samples are removed are down-sampled to 50m x 50m, and sigma is generated0Set of sigma50*50。
Step (32), for sigma0Set of sigma50*50K-means clustering is performed since all pixels σ0The frequency distribution histogram of the values presents mixed Gaussian distribution, the surface of the water body is smooth, the water body has specular reflection characteristics, most microwave signals can be reflected, and therefore the water body has a lower backscattering value compared with other ground objects and is usually dark black on an SAR image. Therefore, according to the feature and mixture of the ground featuresThe Gaussian distribution can divide the ground objects into water bodies and non-water bodies, sigma0Water body class, sigma, is lower in value0The higher value is a non-water body class;
step (33) of calculating σ0The ratio of the number of pixels in the water body pixel set with a lower value to the number of all pixels is the initial value of the prior probability of the water body, namely p0(W)。
In the step (4), the maximum Expectation (EM) algorithm and the water body prior probability in the step (3) are utilized to carry out down-sampling on the set sigma50*50Performing iterative computation distribution parameters to improve the iterative speed, and then performing water body probability computation of each 10m × 10m pixel, specifically comprising the following substeps:
step (41), constructing sigma by using mixed Gaussian distribution0A probability density function of the marginal distribution;
and (42) carrying out iterative optimization on p (W) by utilizing a maximum Expectation (EM) algorithm based on maximum likelihood estimation to carry out iterative calculation on the mean value and standard deviation parameters of the distribution curve.
A step (43) of converting p (W) optimized in the step (42),
Figure BDA0003135844920000081
p(σ0I W) and
Figure BDA0003135844920000082
substituting into a water probability calculation formula;
step (44) of converting the sigma of the two time phase images0Substituting into a formula, respectively calculating the water body probability of the two time phase images (as shown in fig. 3), wherein the left side and the right side are respectively the water body probability values of the region of 26 days in 6 months in 2020 and 7 months in 2020, the probability value threshold is 0-1, the closer to 1, the higher the probability that the pixel belongs to the water body, and the smaller the probability is vice versa.
In the step (5), the water body probabilities of two time-phase images before and after the flood event occurs are differenced, the water body probability increment boundary values of the flood risk area and the non-risk area are calculated, and the result is expressed in the form of water body probability increment, and the method specifically comprises the following substeps:
step (51), calculating the difference of the water body probabilities of the front two time phase images after the flood event;
and (52) taking 0 as a key threshold value of the difference of the water body probabilities of the two time-phase images, extracting the flood risk area by using the water body probabilities, and expressing the flood risk area in a probability increment mode (as shown in figure 4), wherein theoretically the threshold value of the difference of the water body probabilities of the two time phases is 0-1, and the closer to 1 indicates that the probability of the pixel changing from land to water body is higher.
And (53) further performing threshold segmentation on the Y, setting 0.2, 0.4, 0.6 and 0.8 as thresholds, wherein 0-0.2 corresponds to a first level of risk, 0.2-0.4 corresponds to a second level of risk, 0.4-0.6 corresponds to a third level of risk, 0.6-0.8 corresponds to a fourth level of risk, and 0.8-1.0 corresponds to a fifth level of risk.

Claims (7)

1. An automatic flood risk area extraction method based on SAR images is characterized by comprising the following steps:
step (1), acquiring two time-phase SAR images before and after an flood event in an area, preprocessing the two time-phase SAR images, and forming a backscattering coefficient (sigma) by all pixels0) Sigma set1；
Step (2), for the set sigma1Drawing a frequency histogram, and assuming that sigma of two-time phase SAR images distributed with water body0Sigma set1According to the Gaussian mixture distribution and for sigma0Sigma set1Filtering the outliers of (1);
step (3) of resampling the two time-phase SAR images filtered in the step (2) with reduced spatial resolution to generate sigma0Sigma set2And assumes the down-sampled sigma2In a mixed gaussian distribution, and the resulting distribution can be used to downsample the pre-set sigma1The probability of the water body is calculated, and the sigma after the down sampling is calculated0Sigma set2Performing k-means clustering, and calculating the prior probability of the water body and the non-water body;
step (4), utilizing the maximum Expectation (EM) algorithm and the water body prior probability in the step (3) to carry out down-sampling on the set sigma2Performing iterative calculation on distribution parameters to improve the iterative speed, and then performing water body probability meter of each pixel by using the two-time-phase SAR image in the step (1)Calculating;
and (5) expressing the change value of the probability of each pixel belonging to the water body before and after the event by using the water body probability increment, determining the water body probability increment boundary value of the flood risk area and the non-risk area, and constructing a flood risk area extraction method based on the two-time phase water body probability.
2. The method for automatically extracting the flood risk area based on the SAR image according to claim 1, wherein the method comprises the following steps: the step (1) comprises the following specific sub-steps:
step (11), acquiring SAR images covering a research area before and after a flooding event;
step (12), the two time phase SAR images are subjected to registration, multi-time phase filtering, geocoding, radiometric calibration and spatial cutting to obtain sigma0(unit: dB);
step (13), all pixels of the SAR image in two time phases are combined to form sigma0Sigma set1The spatial resolution is the spatial resolution of the two-time phase SAR image.
3. The method for automatically extracting the flood risk area based on the SAR image according to claim 1, wherein the method comprises the following steps: the step (2) comprises the following specific sub-steps:
step (21) for σ0Sigma set1Drawing a frequency histogram, and assuming that a probability density curve of the frequency histogram accords with mixed Gaussian distribution;
and (22) even if a series of preprocessing is carried out on the SAR image by using the terrain data and the two-time phase characteristics, a larger abnormal value still exists to influence the subsequent parameter estimation and the water body probability calculation, so that outliers in the frequency histogram are removed.
4. The method for automatically extracting the flood risk area based on the SAR image according to claim 1, wherein the method comprises the following steps: the step (3) comprises the following specific sub-steps:
step (31), in order to increase the iterative computation speed in step (4), two corresponding pairs after filtering in step (2) are performedResampling the time phase SAR image to reduce spatial resolution and generate sigma0Sigma set2And assume the set sigma2Is also in accordance with a Gaussian mixture distribution and is based on the set sigma2Can be applied to the sigma of the spatial resolution of the step (1)0And calculating the probability of the water body.
Step (32), for sigma0Sigma set2K mean value clustering is carried out, and sigma of water body class is collected0Sigma of aggregate and non-water classes0Gathering;
step (33), calculating the prior probability of the water body and the non-water body; the ratio of the number of the pixels in the water body pixel set to the number of all the pixels is the prior probability p of the water body0(W) and p0(W) is an initial value of p (W), the ratio of the number of the pixels of the non-water body pixel set to the number of all the pixels is a non-water body prior probability, and the formula is as follows:
Figure FDA0003135844910000021
wherein W is a water body pixel set,
Figure FDA0003135844910000022
is a non-aqueous body pixel set, p (W) and
Figure FDA0003135844910000023
respectively representing the water body prior probability and the non-water body prior probability, and the sum of the two is equal to 1.
5. The method for automatically extracting the flood risk area based on the SAR image according to claim 1, wherein the method comprises the following steps: the step (4) comprises the following specific sub-steps:
step (41), constructing sigma by using mixed Gaussian distribution0A probability density function of the marginal distribution, the density function being as follows:
Figure FDA0003135844910000024
wherein W is a water body pixel set,
Figure FDA0003135844910000031
is a non-aqueous body pixel set, p (W) and
Figure FDA0003135844910000032
respectively representing the prior probability of water and the prior probability of non-water, p (sigma)0) Representing the probability that the pixel corresponding to the backscattering coefficient belongs to the water body;
Figure FDA0003135844910000033
Figure FDA0003135844910000034
wherein, muW、sW、
Figure FDA0003135844910000035
Are respectively water body pixel sigma0Mean value, standard deviation and non-water body pixel sigma0Mean, standard deviation of;
and (42) iteratively calculating the mean value and standard deviation parameters of the distribution curve by using a maximum Expectation (EM) algorithm based on maximum likelihood estimation, and iteratively optimizing p (W), wherein the formula is as follows:
Figure FDA0003135844910000036
wherein EM (DEG) represents the maximum Expectation (EM) algorithm, p (W) is the water body prior probability obtained by k-means clustering, sigma0As backscatter coefficient data,. mu.W、sW、
Figure FDA0003135844910000037
Are respectively water body pixel sigma0Mean value, standard deviation and non-water body pixel sigma0Mean, standard deviation of;
a step (43) of converting p (W) optimized in the step (42),
Figure FDA0003135844910000038
p(σ0I W) and
Figure FDA0003135844910000039
substituting the water body probability calculation formula to construct a water body probability calculation formula, wherein the formula is as follows:
p(W|σ0)＝p(W)p(σ0|W)/p(σ0) (7)
step (44) of converting the sigma of the two time phase images0Substituting into a formula, and respectively calculating the water body probability of the two time-phase SAR images.
6. The method for automatically extracting the flooding risk area based on the SAR image as claimed in claim 1, wherein the step (5) comprises the following specific sub-steps:
step (51), the difference of the water body probabilities before and after the flood event is obtained, wherein the water body probability before the flood event is recorded as p1And the probability of the water body after the flood event is recorded as p2Subtracting the water body probability value before the flood event from the water body probability value after the flood event, and recording as p21；
In step (52), the backscattering is rough due to wind and the like, so that the backscattering coefficient is increased, but the water body before the rainfall event is always kept as the water body after the rainfall event under the normal condition, and p is ideally kept21There are no negative values, so 0 is taken to be p21The method comprises the following steps of (1) extracting a flood risk area by using water body probability, and expressing the flood risk area in a probability increment mode, wherein the extraction rule is as follows:
Figure FDA0003135844910000041
the area corresponding to the Y being larger than 0 is the flood risk area, the closer the Y is to 1, the more obvious the change of the water body submerging condition is, and the area with the Y being equal to 0 is the non-flood risk area;
and a step (53) of further performing threshold division on Y to set a plurality of thresholds and dividing the risk into a plurality of stages, wherein the closer the Y value is to 1, the greater the risk.
7. The method as claimed in claim 6, wherein in step (53), the risk is divided into five levels, and the thresholds 0.2, 0.4, 0.6 and 0.8 are set, 0-0.2 corresponding to the first level of risk, 0.2-0.4 corresponding to the second level of risk, 0.4-0.6 corresponding to the third level of risk, 0.6-0.8 corresponding to the fourth level of risk, and 0.8-1.0 corresponding to the fifth level of risk.
CN202110718205.3A 2021-06-28 2021-06-28 Automatic extraction method for flood risk area based on SAR image Active CN113567981B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110718205.3A CN113567981B (en) 2021-06-28 2021-06-28 Automatic extraction method for flood risk area based on SAR image

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110718205.3A CN113567981B (en) 2021-06-28 2021-06-28 Automatic extraction method for flood risk area based on SAR image

Publications (2)

Publication Number Publication Date
CN113567981A true CN113567981A (en) 2021-10-29
CN113567981B CN113567981B (en) 2023-08-08

Family

ID=78162971

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110718205.3A Active CN113567981B (en) 2021-06-28 2021-06-28 Automatic extraction method for flood risk area based on SAR image

Country Status (1)

Country Link
CN (1) CN113567981B (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113920448A (en) * 2021-12-15 2022-01-11 航天宏图信息技术股份有限公司 Flood inundation information extraction method and device, electronic equipment and storage medium
CN114067152A (en) * 2022-01-14 2022-02-18 南湖实验室 Refined flood inundated area extraction method based on satellite-borne SAR image
CN116630426A (en) * 2023-07-21 2023-08-22 海南卫星海洋应用研究院有限公司 Flood inundation area extraction method and system
CN117274821A (en) * 2023-11-20 2023-12-22 中国石油大学(华东) Multi-polarization SAR farmland flood detection method and system considering rainfall influence
CN117310705A (en) * 2023-11-28 2023-12-29 中国石油大学(华东) Flood disaster rapid detection method based on dual-polarized SAR image

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CA2787896A1 (en) * 2010-02-01 2011-08-04 Michael Eineder A method for measuring the water level of a body of water
CN102999897A (en) * 2011-09-19 2013-03-27 香港中文大学 Method and device for sea surface oil spillage detection based on SAR (synthetic aperture radar) image
CN103984945A (en) * 2014-05-14 2014-08-13 武汉大学 Optical remote sensing image ship detection method
WO2016024050A1 (en) * 2014-08-12 2016-02-18 Total Sa Method for detecting hydrocarbon deposits
CN106778629A (en) * 2016-12-21 2017-05-31 中国科学院深圳先进技术研究院 Greenhouse recognition methods and device
CN108932520A (en) * 2018-04-26 2018-12-04 武汉大学 In conjunction with the SAR image water body probability drafting method of prior probably estimation
CN110781775A (en) * 2019-10-10 2020-02-11 武汉大学 Remote sensing image water body information accurate segmentation method supported by multi-scale features
CN111259876A (en) * 2020-05-06 2020-06-09 中国水利水电科学研究院 Radar data water body information extraction method and system based on land surface water body product
CN112030909A (en) * 2020-08-21 2020-12-04 江苏东方生态清淤工程有限公司 Full-flow all-season treatment process for controlling lake flooding of lake water body
CN112700431A (en) * 2021-01-11 2021-04-23 武汉大学 Flood coverage surface extraction method based on remote sensing image

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CA2787896A1 (en) * 2010-02-01 2011-08-04 Michael Eineder A method for measuring the water level of a body of water
CN102999897A (en) * 2011-09-19 2013-03-27 香港中文大学 Method and device for sea surface oil spillage detection based on SAR (synthetic aperture radar) image
CN103984945A (en) * 2014-05-14 2014-08-13 武汉大学 Optical remote sensing image ship detection method
WO2016024050A1 (en) * 2014-08-12 2016-02-18 Total Sa Method for detecting hydrocarbon deposits
CN106778629A (en) * 2016-12-21 2017-05-31 中国科学院深圳先进技术研究院 Greenhouse recognition methods and device
CN108932520A (en) * 2018-04-26 2018-12-04 武汉大学 In conjunction with the SAR image water body probability drafting method of prior probably estimation
CN110781775A (en) * 2019-10-10 2020-02-11 武汉大学 Remote sensing image water body information accurate segmentation method supported by multi-scale features
CN111259876A (en) * 2020-05-06 2020-06-09 中国水利水电科学研究院 Radar data water body information extraction method and system based on land surface water body product
CN112030909A (en) * 2020-08-21 2020-12-04 江苏东方生态清淤工程有限公司 Full-flow all-season treatment process for controlling lake flooding of lake water body
CN112700431A (en) * 2021-01-11 2021-04-23 武汉大学 Flood coverage surface extraction method based on remote sensing image

Non-Patent Citations (8)

* Cited by examiner, † Cited by third party
Title
LAURA GIUSTARINI 等: "Probabilistic Flood Mapping Using Synthetic Aperture Radar Data", 《IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING》, pages 6958 - 6969 *
周雨婷 等: "基于改进BP神经网络的非平稳水文时间序列长期预报研究", 《2020年（第八届）中国水利信息化技术论坛论文集》, pages 85 - 92 *
孙忠华 等: "空间信息技术在洪涝险情分析中的应用实践", 《人民长江》, pages 196 - 202 *
孟令奎 等: "遥感影像水体提取与洪水监测应用综述", 《水利信息化》, pages 339 - 447 *
张伟 等: "一种利用多时相GF-4影像的快速水体提取方法", 《遥感信息》, pages 108 - 114 *
湛南渝: ""台风-暴雨"洪涝灾害遥感监测与评估研究", 《中国优秀硕士学位论文全文数据库 基础科学辑》, pages 1 - 85 *
罗固源 等: "邻苯二甲酸酯在长江重庆段水体的概率风险分析", 《长江流域资源与环境》, pages 79 - 83 *
谷鑫志 等: "高分三号影像水体信息提取", 《遥感学报》, pages 555 - 565 *

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113920448A (en) * 2021-12-15 2022-01-11 航天宏图信息技术股份有限公司 Flood inundation information extraction method and device, electronic equipment and storage medium
CN113920448B (en) * 2021-12-15 2022-03-08 航天宏图信息技术股份有限公司 Flood inundation information extraction method and device, electronic equipment and storage medium
CN114067152A (en) * 2022-01-14 2022-02-18 南湖实验室 Refined flood inundated area extraction method based on satellite-borne SAR image
CN116630426A (en) * 2023-07-21 2023-08-22 海南卫星海洋应用研究院有限公司 Flood inundation area extraction method and system
CN117274821A (en) * 2023-11-20 2023-12-22 中国石油大学(华东) Multi-polarization SAR farmland flood detection method and system considering rainfall influence
CN117274821B (en) * 2023-11-20 2024-02-02 中国石油大学(华东) Multi-polarization SAR farmland flood detection method and system considering rainfall influence
CN117310705A (en) * 2023-11-28 2023-12-29 中国石油大学(华东) Flood disaster rapid detection method based on dual-polarized SAR image
CN117310705B (en) * 2023-11-28 2024-02-09 中国石油大学(华东) Flood disaster rapid detection method based on dual-polarized SAR image

Also Published As

Publication number Publication date
CN113567981B (en) 2023-08-08

Similar Documents

Publication Publication Date Title
CN113567981B (en) Automatic extraction method for flood risk area based on SAR image
CN109977801B (en) Optical and radar combined regional water body rapid dynamic extraction method and system
Li et al. The recent developments in cloud removal approaches of MODIS snow cover product
CN112800634B (en) Rainfall estimation method and system coupling dry-wet state identification and multi-source information fusion
CN111008941A (en) Agricultural flood disaster range monitoring system and method based on high-resolution satellite remote sensing image
Greifeneder et al. Suitability of SAR imagery for automatic flood mapping in the Lower Mekong Basin
CN112084712B (en) Flood inundation range dynamic simulation method integrating active and passive microwave remote sensing information
Wu et al. Cloud-based typhoon-derived paddy rice flooding and lodging detection using multi-temporal Sentinel-1&2
Wan et al. Automatic extraction of flood inundation areas from SAR images: A case study of Jilin, China during the 2017 flood disaster
Zhu et al. An advanced change detection method for time-series soil moisture retrieval from Sentinel-1
Ai et al. The impact of rapid urban expansion on coastal mangroves: a case study in Guangdong Province, China
CN114998213B (en) Runoff weight model-based lake and reservoir flood retention amount calculation method
CN116704331A (en) Beach change information extraction method and device and electronic equipment
Gašparović et al. Green infrastructure mapping in urban areas using sentinel-1 imagery
CN112967308B (en) Amphibious boundary extraction method and system for dual-polarized SAR image
Thissen Automating Surface Water Detection for Rivers
Zhu et al. A XGBoost-based downscaling-calibration scheme for extreme precipitation events
Kadapala et al. Region-growing-based automatic localized adaptive thresholding algorithm for water extraction using sentinel-2 MSI imagery
Zhang et al. Comparative strengths of SCaMPR satellite QPEs with and without TRMM ingest versus gridded gauge-only analyses
Qi et al. An adaptive threshold selected method from remote sensing image based on water index
Dou et al. Spatio-temporal evolution of glacial lakes in the Tibetan Plateau over the past 30 years
Huang et al. Image similarity-based gap filling method can effectively enrich surface water mapping information
Yen et al. Coastline changes detection from Sentinel–1 satellite imagery using spatial fuzzy clustering and interactive thresholding method in Phan Thiet, Binh Thuan
Zhu et al. A Coupled Temporal–Spectral–Spatial Multidimensional Information Change Detection Framework Method: A Case of the 1990–2020 Tianjin, China
CN115294183A (en) Disc-shaped sub-lake water body time sequence extraction method based on multi-source remote sensing 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