CN112967308A - Amphibious SAR image boundary extraction method and system - Google Patents
Amphibious SAR image boundary extraction method and system Download PDFInfo
- Publication number
- CN112967308A CN112967308A CN202110220619.3A CN202110220619A CN112967308A CN 112967308 A CN112967308 A CN 112967308A CN 202110220619 A CN202110220619 A CN 202110220619A CN 112967308 A CN112967308 A CN 112967308A
- Authority
- CN
- China
- Prior art keywords
- data
- polarized
- land
- dual
- value
- 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
- 238000000605 extraction Methods 0.000 title claims description 46
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims abstract description 102
- 238000000034 method Methods 0.000 claims abstract description 29
- 230000010287 polarization Effects 0.000 claims description 37
- 230000003287 optical effect Effects 0.000 claims description 33
- 238000010586 diagram Methods 0.000 claims description 28
- 238000012545 processing Methods 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 8
- 238000006243 chemical reaction Methods 0.000 claims description 5
- 238000007405 data analysis Methods 0.000 claims description 4
- 230000000903 blocking effect Effects 0.000 claims description 3
- 238000013075 data extraction Methods 0.000 claims description 3
- 230000001172 regenerating effect Effects 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 2
- 230000009977 dual effect Effects 0.000 claims 1
- 238000009826 distribution Methods 0.000 description 6
- 238000003384 imaging method Methods 0.000 description 4
- 238000012937 correction Methods 0.000 description 3
- 230000005855 radiation Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000002902 bimodal effect Effects 0.000 description 2
- 230000001427 coherent effect Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000008602 contraction Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000002329 infrared spectrum Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000002310 reflectometry Methods 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
- 238000003809 water extraction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/181—Segmentation; Edge detection involving edge growing; involving edge linking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/40—Image enhancement or restoration using histogram techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
- G06T2207/30184—Infrastructure
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Cheminformatics (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)
- Geometry (AREA)
- Image Processing (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention discloses a method and a system for extracting land and water boundaries of a dual-polarized SAR image.
Description
Technical Field
The invention relates to the technical field of computer image processing, in particular to a method and a system for extracting an amphibious boundary of a dual-polarized SAR image.
Background
The influence of climate change increases the frequency and probability of extreme weather, which is expressed by flood and drought disasters. The automatic high-precision extraction of the large-area land-water boundary is beneficial to the real-time or quasi-real-time monitoring/early warning of flood disasters and drought disasters. The following three main land and water boundary extraction technologies exist:
and (5) extracting the water and land boundary by using the optical remote sensing image. This type of image has inherent deficiencies, and is very susceptible to adverse weather such as rain, snow, cloud, haze and the like, due to the fact that radiation from visible light to near infrared spectrum is difficult to penetrate through cloud layers or dense fog, and is particularly serious in cloudy and rainy regions in the south. Because the flood disaster is often accompanied by cloud and rain weather, the optical remote sensing image is difficult to play a role in the important application scene.
Single polarization SAR image bimodal method. The characteristic that a water body and a land on a single-polarized SAR image have a double-peak structure is utilized, and a valley value between double peaks is used as a threshold value of density segmentation, so that land and water boundaries are extracted. The accuracy of this method is questionable because the so-called bimodal structure is actually the result of the superposition of two normal distributions of water and land, the intersection of which makes a considerable fraction of the misdistribution errors, whatever the threshold value used. The real land is wrongly divided into the land or the real land is wrongly divided into the water at the land-water boundary. The information content of the single-polarized SAR image is limited, and the error can be reduced as much as possible by using the valley value between the two peaks as a threshold value.
And (3) a dual-polarization SAR image water body index method. The SAR image of two polarization modes is comprehensively applied by introducing the concept of water body index, so that the water body identification precision is further improved. The calculation formula of the water body index is as follows: wi ═ ln (10 × p)1*p2) Wherein p1 and p2 are backscattering coefficients or decibel-normalized intensity values in two polarization modes, respectively. And then on the basis of the water body index, performing extraction processing like a single polarization double peak method to obtain the water-land boundary. The method utilizes more abundant information under two polarization modes, and can obtain a result more accurate than that of single polarization, but the used method and the single polarization double-peak method are not two-fold, and the information of double polarization is not utilized to the maximum extent.
Disclosure of Invention
In order to solve the problems, the invention provides a method and a system for extracting the land and water boundaries of a dual-polarized SAR image, and aims to fully explore rich information contained in the dual-polarized SAR image so as to accurately define the land and water boundaries in an image.
In order to solve the technical problem, a first aspect of the present invention provides a method for extracting an amphibious boundary of a dual-polarized SAR image, including the following steps:
s1, acquiring dual-polarized SAR image data of the target area;
s2, the two types of single-polarized SAR image data contained in the dual-polarized SAR image data are counted to respectively obtain frequency histograms of the first type of polarized data and the second type of polarized data, and a first water body peak value w is extracted from the frequency histogram of the first type of polarized data1First land peak l1And a first valley value v1Extracting a second water body peak value w from the frequency histogram of the second type polarization data2Second land peak l2And a second valley value v2；
S3, establishing a Cartesian coordinate system by taking the first type of polarization data as an X axis and the second type of polarization data as a Y axis, and marking W (W)1，w2)、L(l1，l2) And V (V)1，v2) Three feature points;
s4, drawing a straight line k through the W point and the L point1Passing the V point perpendicular to the line k1Straight line k of2Falls on the straight line k2The dual-polarized data point at the lower left is divided into water bodies and falls on the straight line k2The upper right dual-polarized data points are divided into lands, and a single-time phase amphibian map extraction result is generated.
In some embodiments, the S2 is specifically:
s21, blocking the first type of polarization data/the second type of polarization data in units of preset pixels to form a grid formed by M × N image blocks, and merging pixels that are less than the preset pixels into a last image block;
s22, removing background values in the image blocks to obtain effective values, and extracting histogram peak-valley values of the effective values one by one;
s23, randomly extracting 10% of pixels in the effective value to form a set S;
s24, sorting the data of the set S from small to large, taking the 1% quantile of the sorted set S as the minimum value of the data range, and taking the 99% quantile of the sorted set S as the maximum value of the data range;
s25, equally dividing the data range into 256 intervals, and calculating the number of pixels falling into each interval to obtain a frequency histogram of the first type of polarized data/the second type of polarized data;
s26, traversing 256 intervals of the frequency histogram, and if the pixel value of the current target interval appearing for the first time is the maximum value in the 9 nearest neighbor intervals, defining the target interval as the first water body peak value w1The second water body peak value w2If the pixel value of the current target interval appears for the second time is the maximumClose to the maximum value in 9 intervals, defining the target interval as the first land peak value l1The second terrestrial peak/2；
S27, locating the peak value w of the first water body1And said first terrestrial peak/1The second water body peak value w2And said second terrestrial peak/2The interval of the minimum value in all the intervals in between is recorded as the first valley value v1The second valley value v2。
In some embodiments, the method further includes performing secondary classification if the number difference between the water body pixels and the land pixels included in the single-phase amphibian map extraction result is more than twice.
In some embodiments, the secondary classification includes randomly extracting 10% of pixels of a smaller number between the water body pixels and the land pixels from the water body range and the land range of the single-phase amphibian map extraction result, taking a union of the pixels extracted from the water body range and the land range as a secondary classification calculation object, and repeating the steps of S24-S27 and S3-S4 to regenerate the single-phase amphibian map extraction result.
In some embodiments, the method further includes obtaining optical image data closest to the dual-polarization SAR image data in time, extracting an optical amphibian map result from the optical image data by using an optical remote sensing image vegetation and water body automatic extraction method, defining an intersection of the single-time phase amphibian map extraction result and the optical amphibian map result as a combined binary map result, and performing raster vector conversion processing on the single-time phase amphibian map extraction result, the optical amphibian map result and the combined binary map result to obtain three vector maps V respectively1、V2And V3Obtaining a vector image V having an intersection relation with the vector image V31A vector diagram V' formed by all the entities in (A)1。
In some embodiments, further comprising, for vector image V3Each entity in (2) performs the following processing: is provided withThe current entity is E3Computing entity E3Has an area of A3Calculating to obtain the sum entity E3Vector diagram V' with intersection relation1Corresponding entity E in1Has an area of A1Calculating to obtain the sum entity E3Vector diagram V with intersection relation2Corresponding entity E in2Has an area of A2Definition of A1And A2The smaller value therebetween is AsDefinition of A1And A2The larger value in between is AbDefinition of degree of inclusion INC ═ A3/AsThe degree of expansion is defined as EXP ═ Ab/AsIf INC<0.8 or EXP>500%, entity E1From vector diagram V1Is deleted.
In some embodiments, after acquiring the dual-polarized SAR image data of the target region, the dual-polarized SAR image data is preprocessed.
Meanwhile, the invention provides an amphibious boundary extraction system of a dual-polarized SAR image in a second aspect, which comprises:
the data acquisition device is used for acquiring dual-polarized SAR image data of a target area;
a data analysis device for counting two types of single-polarized SAR image data contained in the dual-polarized SAR image data, respectively obtaining frequency histograms of the first type of polarized data and the second type of polarized data, and extracting a first water peak value w from the frequency histogram of the first type of polarized data1First land peak l1And a first valley value v1Extracting a second water body peak value w from the frequency histogram of the second type polarization data2Second land peak l2And a second valley value v2；
The data modeling device is used for establishing a Cartesian coordinate system by taking the first type of polarization data as an X axis and the second type of polarization data as a Y axis and marking W (W)1，w2)、L(l1，l2) And V (V)1，v2) Three feature points;
data extraction means for making a straight line passing through the W point and the L pointk1Passing the V point perpendicular to the line k1Straight line k of2Falls on the straight line k2The dual-polarized data point at the lower left is divided into water bodies and falls on the straight line k2The upper right dual-polarized data points are divided into lands, and a single-time phase amphibian map extraction result is generated.
The invention has the beneficial effects that: the method comprises the steps of extracting peak-valley values from a frequency histogram and establishing a two-dimensional scatter coordinate system by counting the frequency histogram of dual-polarized SAR image data, generating a single-time-phase amphibian map extraction result by taking a straight line as an amphibian discriminant of dual-polarized data scatter points, and obtaining an accurate and reliable amphibian boundary, wherein the automation degree is high, and the whole process does not need any manual operation.
Drawings
Fig. 1 is a schematic flow chart of a disclosed method for extracting an amphibious boundary of a dual-polarized SAR image according to an embodiment of the present invention;
fig. 2a is a dual-polarized SAR image of an east river lake region of chenzhou city, hunan, 12 months and 9 days in 2020;
fig. 2b is a binary image of the land and water boundary extraction result of the single-time-phase dual-polarized SAR image;
FIG. 3a is an optical image of the east Jianghu area of Chenzhou city, Hunan, 11 months and 10 days in 2020;
FIG. 3b is a binary image of the optical land-water boundary result of the optical image;
fig. 4a is an amphibious vector diagram of the east river lake region of 12-9-2020-12-9 days after shadow removal;
FIG. 4b is an amphibian map of the east river lake region at 12/9/2020 after shadow removal;
fig. 5 is a schematic structural diagram of an amphibious boundary extraction system of a dual-polarized SAR image according to a third embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer and clearer, the following description will use a dual-polarized SAR image of a sentinel one # in the east river lake region of chenzhou city, hunjiang, 9 th of 2020 by 12 th of 2020 as an example with reference to the accompanying drawings and the specific implementation manner. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some but not all of the relevant aspects of the present invention are shown in the drawings.
Example one
As shown in fig. 1, the present embodiment provides a method for extracting an amphibious boundary of a dual-polarized SAR image, including the following steps:
and S1, acquiring dual-polarized SAR image data of the target area.
In this embodiment, the target area is selected as the east river lake area of Chen city, Hunan province, and dual-polarized SAR image data provided by Sentinel one (Sentinel-1) is downloaded from the office of the European Space Administration (ESA) through a PC or any network-enabled device, the website of the dual-polarized SAR image data is https:// scihub. copernicus. eu/dhus/#/home, the time of the dual-polarized SAR image data file is 12 months and 9 days 2020, the file name is S1A _ IW _ GRDH _1SDV _20201209T103450_20201209T103515_035608_042A35_8945.zip, the product type is GRD, and the product has two polarization modes of VV and VH.
Furthermore, after acquiring the dual-polarization SAR image data of the target area, preprocessing the dual-polarization SAR image data, which mainly comprises the following steps: 1) track correction, namely automatically downloading an accurate track file and updating the state information of the Sentinel-1 satellite track in the xml metadata file; 2) thermal noise removal, wherein the thermal noise of the SAR image can influence the accuracy of the radar backscattering signal, and the signal-to-noise ratio of the SAR image can be improved by performing the thermal noise removal; 3) radiometric calibration, which converts the backscattering signal received by the sensor into backscattering coefficient; 4) coherent speckle filtering, which is to use a referred-Lee filter (an adaptive filter) to filter coherent speckle noise in an image; 5) the method comprises the following steps of terrain correction, namely automatically downloading a DEM (digital elevation model) to perform terrain correction on an image and correcting position deviation caused by terrain fluctuation; 6) decibel, which is essentially a logarithmic transformation, has a backscattering coefficient that is approximately a common gaussian distribution after decibel conversion, and is favorable for visualization and data analysis.
The preprocessed 12-month-2020-9-day dual-polarized SAR image data is shown in fig. 2a, and since the algorithm flow of each image block is completely the same, the example only intercepts 5000-pixel square image blocks including the range of the east river lake as an example.
S2, the two types of single-polarized SAR image data contained in the dual-polarized SAR image data are counted to respectively obtain the frequency histograms of the first type of polarized data and the second type of polarized data, and the first water peak value w is extracted from the frequency histogram of the first type of polarized data1First land peak l1And a first valley value v1Extracting a second water body peak value w from the frequency histogram of the second type polarization data2Second land peak l2And a second valley value v2；
The above S2 specifically includes:
s21, blocking the first type polarization data/the second type polarization data in units of preset pixels to form a grid formed by M × N image blocks, and merging pixels less than the preset pixels into the last image block;
in this embodiment, the first type polarization data/the second type polarization data are not blocked, i.e. there is no 1 × 1 original image.
S22, removing the background value (generally 0) in the image block to obtain effective values, and extracting the histogram peak-valley values of the effective values one by one;
and S23, randomly extracting 10% of pixels in the effective value to form a set S, and using the set S as a primary classification calculation object to enable the calculation speed to be higher. Because the statistical characteristics of the random samples are utilized to represent the global statistical characteristics, 10% of samples are randomly extracted to calculate the characteristic value, and the speed is increased by 10 times. Taking the sentinel one satellite as an example, the land and water boundary extraction in an area of 300 kilometers by 200 kilometers requires less than 10 minutes.
S24, in order to eliminate the interference of extreme abnormal values, sorting the data of the set S from small to large, taking the 1% quantile of the sorted set S as the minimum value of the data range, and taking the 99% quantile of the sorted set S as the maximum value of the data range;
s25, equally dividing the data range into 256 intervals, and calculating the number of pixels falling into each interval to obtain a frequency histogram of the first type polarization data/the second type polarization data;
s26, traversing 256 intervals, and if the pixel value of the current target interval appears for the first time and is the maximum value of 9 nearest neighbor intervals (including the interval and four intervals at the left and right), defining the target interval as a first water body peak value w1Second water body peak value w2If the pixel value of the current target interval appearing for the second time is the maximum value in the 9 nearest neighbor intervals, defining the target interval as a first land peak value l1Second land Peak l2；
S27, locating the peak value w of the first water body1And a first land peak l1Second water body peak value w2And a second land peak/2The minimum value in all the intervals in between is recorded as the first valley value v1Second trough value v2。
S3, establishing a Cartesian coordinate system with the first type of polarization data as an X axis and the second type of polarization data as a Y axis, and marking W (W)1，w2)、L(l1，l2) And V (V)1，v2) And the three characteristic points respectively represent scattered points with the maximum water body distribution density, the maximum land distribution density and the minimum land-water distribution density.
S4, drawing a straight line k through the W point and the L point1Passing V point perpendicular to line k1Straight line k of2In a straight line k2Is used as the discriminant of water and land and falls on the straight line k2The left lower dual-polarized data point is divided into water bodies and falls on a straight line k2The upper right dual-polarized data points are divided into lands, and a single-time phase land and water boundary extraction result is generated, wherein the result is actually a binary image (the value of the water body is 1, and the value of the land is 0) and is marked as B1As shown in fig. 2 b.
According to the method, the peak-valley value is extracted from the frequency histogram and a two-dimensional scatter coordinate system is established by counting the frequency histogram of the dual-polarized SAR image data, the single-time-phase amphibian map extraction result is generated by taking the straight line as the amphibian discriminant of the dual-polarized data scatter, the accurate and reliable amphibian boundary is obtained, the automation degree is high, and no manual operation is needed in the whole process.
In order to enhance the classification accuracy, the single-time phase amphibian map extraction result can be processed by adopting the following steps: and if the quantity difference between the water body pixels and the land pixels contained in the single-time-phase land-water binary image extraction result is more than twice, performing secondary classification. And the secondary classification comprises the steps of randomly extracting 10% of pixels with the smaller quantity between the water body pixels and the land pixels from the water body range and the land range of the single-time-phase amphibian map extraction result, taking the union set of the pixels extracted from the water body range and the land range as a secondary classification calculation object, repeating the steps of S24-S27 and S3-S4, and regenerating the single-time-phase amphibian map extraction result.
In this embodiment, let the number of pixels in the water be NwAnd the number of land pixels is NlTaking the larger value of the two as NbThe smaller value of both is NsIf N is presentb/Ns>2, considering that the land and water proportions are greatly different, and carrying out secondary classification. This is because on the histogram, the dominant one has an annihilation effect on the weak potential side, so that the valley between the peaks moves toward the weak potential side to deviate from a reasonable position. When the dual-polarized valley value deviates, the straight line for discrimination can also translate, thereby influencing the precision of the classification result. The secondary classification steps are as follows: randomly extracting N from the water body range and land range of the primary classificationsAnd taking the union of the 10% of the pixels as a secondary classification calculation object, repeating the steps of S24-S27 and S3-S4, and extracting the result of the single-time-phase amphibian diagram.
Example two
The influence of the shadow on the water body is not well handled by the three land and water boundary extraction methods in the background art. The existence of mountains or tall buildings and the characteristics of satellite remote sensing oblique photogrammetry enable more shadows to exist on optical or radar images, and the problem that the interference of shadow noise is removed as far as possible on the basis of water extraction is always puzzled by people. Because the imaging modes and the imaging mechanisms of the optical satellite and the radar satellite are greatly different, particularly, the optical satellite belongs to passive remote sensing, and the reflectivity of the ground object to solar radiation is recorded. The radar satellite belongs to active remote sensing, and the backscattering coefficient of ground objects to radar active emission radiation is recorded. The imaging positions and observation directions of different satellites at different time are different, and the relative azimuth angles between the sensor and the ground object are also different, so that the shadow of a mountain or a building appears on different sides of the entity, and the shadow of the same object does not intersect completely or has a small degree of inclusion even if the shadow has an intersection relationship in two different observations in a short period (a smaller entity in the two periods is included in a larger entity). The position of the water body is not changed in a short time, and only expansion or contraction occurs at the boundary of the land and the water, so that the water body has a larger inclusion degree. The different points of the water body and the shadow on the time phase characteristics are fully utilized to distinguish the water body and the shadow.
Therefore, in the embodiment, on the basis of the first embodiment, a method for removing shadows by a multi-source multi-temporal method is added to remove radar image shadows mixed in the water body.
The shadow removal steps of the multi-temporal method are as follows: firstly, optical image data which is closest to the dual-polarization SAR image data in terms of time is obtained, and a target area is required to be shielded without a cloud layer.
In this example, the optical image data is the latest first-phase historical sentinel second image data downloaded from the same website, as shown in fig. 3 a. The time stamp of the optical image data is 11 months and 10 days in 2020, and the file name is S2B _ MSIL2A _20201110T025939_ N0214_ R032_ T49RGJ _20201110T055638. zip. The data in the designated area and the time range can be automatically downloaded by selecting manual downloading or calling a function in a sentinelsat library by using Python.
Then, an optical remote sensing image vegetation and water body automatic extraction method is adopted to extract an optical land and water binary image result from the optical image data, and the result is actually a binary image (the value of the water body is 1, and the value of the land is 0) and is recorded as B2As shown in fig. 3 b. Defining the intersection of the single-time phase amphibian map extraction result and the optical amphibian map result as a combined binary map result(i.e. B)1Is 1 # B2Is 1), a further binary image is obtained, denoted as B3Extracting result B from single-time phase amphibian diagram1Optical land-water binary image result B2And merging the binary image results B3Carrying out grid vector conversion processing to respectively obtain three vector diagrams V1、V2And V3Obtaining a vector image V having an intersection relation with the vector image V31Vector diagram V formed by all the entities in (1)1Hitherto, shadows having no intersection relationship at different imaging times have been removed.
Further shadow removal, for vector image V3Each entity in (2) performs the following processing: for vector diagram V3Each entity in (2) performs the following processing: let the current entity be E3Computing entity E3Has an area of A3Calculating to obtain the sum entity E3Vector diagram V' with intersection relation1Corresponding entity E in1Has an area of A1Calculating to obtain the sum entity E3Vector diagram V with intersection relation2Corresponding entity E in2Has an area of A2Definition of A1And A2The smaller value therebetween is AsDefinition of A1And A2The larger value in between is AbDefinition of degree of inclusion INC ═ A3/AsThe degree of expansion is defined as EXP ═ Ab/AsIf INC<0.8 or EXP>500%, entity E1From vector diagram V1And deleting the shadow which has an intersection relation at different moments but has a small degree of inclusion or a large degree of expansion.
After all the entities are processed, the current-stage land and water boundary vector diagram with most shadows removed can be obtained, for example, fig. 4a illustrates a land and water boundary vector distribution diagram of the east river lake region 12/9/2020 after shadow removal, and fig. 4b is a corresponding land and water binary diagram, so that compared with fig. 2b, the land and water boundary with shadows removed is closer to the real environment. In the embodiment, the single-time-phase land-water binary image extraction result is fused with the multi-source multi-time-phase land-water information, so that the interference of radar shadow is removed to the maximum extent, and the land-water boundary with reliable precision is obtained.
EXAMPLE III
As shown in fig. 5, an amphibious boundary extraction system of a dual-polarized SAR image is disclosed, which includes:
the data acquisition device is used for acquiring dual-polarized SAR image data of a target area;
a data analysis device for counting two types of single-polarized SAR image data contained in the dual-polarized SAR image data, respectively obtaining frequency histograms of the first type of polarized data and the second type of polarized data, and extracting a first water peak value w from the frequency histogram of the first type of polarized data1First land peak l1And a first valley value v1Extracting a second water body peak value w from the frequency histogram of the second type polarization data2Second land peak l2And a second valley value v2；
The data modeling device is used for establishing a Cartesian coordinate system by taking the first type of polarization data as an X axis and the second type of polarization data as a Y axis and marking W (W)1，w2)、L(l1，l2) And V (V)1，v2) Three feature points;
data extraction means for drawing a straight line k through the W point and the L point1Passing V point perpendicular to line k1Straight line k of2Falling on a straight line k2The left lower dual-polarized data point is divided into water bodies and falls on a straight line k2Dividing the upper right dual-polarized data points into lands, generating a single-time phase amphibian map extraction result, and recording the result as B1。
In this embodiment, the apparatuses included in the system may be arranged according to an example of an embodiment, such as:
and the secondary classification device is used for randomly extracting 10% of pixels with the smaller quantity between the water body pixels and the land pixels from the water body range and the land range of the single-time-phase amphibian map extraction result, taking the union set of the pixels extracted from the water body range and the land range as a secondary classification calculation object, repeating the steps of S24-S27 and S3-S4, and regenerating the single-time-phase amphibian map extraction result.
As another implementation manner of this embodiment, the step of removing the shadow by the multi-source multi-temporal method described above may be packaged as a physical/virtual device to implement specific functions:
and the optical image data acquisition device is used for acquiring optical image data which is closest to the dual-polarized SAR image data in terms of time.
A first shadow removing device for extracting an optical land-water binary image result from the optical image data by adopting an optical remote sensing image vegetation and water body automatic extraction method, wherein the result is actually a binary image (the value of the water body is 1, and the value of the land is 0) and is marked as B2And defining the intersection of the single-time phase amphibian map extraction result and the optical amphibian map result as a combined binary map result (namely B)1Is 1 # B2Is 1), a further binary image is obtained, denoted as B3Extracting result B from single-time phase amphibian diagram1Optical land-water binary image result B2And merging the binary image results B3Carrying out grid vector conversion processing to respectively obtain three vector diagrams V1、V2And V3Obtaining a vector image V having an intersection relation with the vector image V31A vector diagram V' formed by all the entities in (A)1。
Second shadow removal means for removing shadow of the vector image V3Each entity in (2) performs the following processing: let the current entity be E3Computing entity E3Has an area of A3Calculating to obtain the sum entity E3Vector diagram V with intersection relation1Corresponding entity E in1Has an area of A1Calculating to obtain the sum entity E3Vector diagram V with intersection relation2Corresponding entity E in2Has an area of A2Definition of A1And A2The smaller value therebetween is AsDefinition of A1And A2The larger value in between is AbDefinition of degree of inclusion INC ═ A3/AsThe degree of expansion is defined as EXP ═ Ab/AsIf INC<0.8 or EXP>500%, entity E1From vector diagram V1Is deleted.
The above embodiments are only for illustrating the technical concept and features of the present invention, and the purpose thereof is to enable those skilled in the art to understand the contents of the present invention and implement the present invention accordingly, and not to limit the protection scope of the present invention accordingly. All equivalent changes or modifications made in accordance with the spirit of the present disclosure are intended to be covered by the scope of the present disclosure.
Claims (8)
1. A method for extracting an amphibious boundary of a dual-polarized SAR image is characterized by comprising the following steps:
s1, acquiring dual-polarized SAR image data of the target area;
s2, the two types of single-polarized SAR image data contained in the dual-polarized SAR image data are counted to respectively obtain frequency histograms of the first type of polarized data and the second type of polarized data, and a first water body peak value w is extracted from the frequency histogram of the first type of polarized data1First land peak l1And a first valley value v1Extracting a second water body peak value w from the frequency histogram of the second type polarization data2Second land peak l2And a second valley value v2；
S3, establishing a Cartesian coordinate system by taking the first type of polarization data as an X axis and the second type of polarization data as a Y axis, and marking W (W)1，w2)、L(l1，l2) And V (V)1，v2) Three feature points;
s4, drawing a straight line k through the W point and the L point1Passing the V point perpendicular to the line k1Straight line k of2Falls on the straight line k2The dual-polarized data point at the lower left is divided into water bodies and falls on the straight line k2The upper right dual-polarized data points are divided into lands, and a single-time phase amphibian map extraction result is generated.
2. The amphibious SAR image land-water boundary extraction method of claim 1, wherein the S2 specifically is as follows:
s21, blocking the first type of polarization data/the second type of polarization data in units of preset pixels to form a grid formed by M × N image blocks, and merging pixels that are less than the preset pixels into a last image block;
s22, removing background values in the image blocks to obtain effective values, and extracting histogram peak-valley values of the effective values one by one;
s23, randomly extracting 10% of pixels in the effective value to form a set S;
s24, sorting the data of the set S from small to large, taking the 1% quantile of the sorted set S as the minimum value of the data range, and taking the 99% quantile of the sorted set S as the maximum value of the data range;
s25, equally dividing the data range into 256 intervals, and calculating the number of pixels falling into each interval to obtain a frequency histogram of the first type of polarized data/the second type of polarized data;
s26, traversing 256 intervals of the frequency histogram, and if the pixel value of the current target interval appearing for the first time is the maximum value in the 9 nearest neighbor intervals, defining the target interval as the first water body peak value w1The second water body peak value w2If the pixel value of the current target interval appearing for the second time is the maximum value in the 9 nearest neighbor intervals, defining the target interval as the first land peak value l1The second terrestrial peak/2；
S27, locating the peak value w of the first water body1And said first terrestrial peak/1The second water body peak value w2And said second terrestrial peak/2The interval of the minimum value in all the intervals in between is recorded as the first valley value v1The second valley value v2。
3. The method for extracting the amphibious boundary of the dual-polarized SAR image according to claim 2, further comprising performing secondary classification if the number difference between the water body pixels and the land pixels included in the single-phase amphibious binary image extraction result is more than twice.
4. The amphibious SAR image land-water boundary extraction method according to claim 3, wherein the secondary classification includes randomly extracting 10% of pixels with a smaller number between the water body pixels and the land pixels from the water body range and the land range of the single-phase land-water binary image extraction result, using the union set of the pixels extracted from the water body range and the land range as a secondary classification calculation object, repeating the steps of S24-S27 and S3-S4, and regenerating the single-phase land-water binary image extraction result.
5. The method for extracting the amphibious SAR image boundary as claimed in claim 1 or 4, further comprising obtaining the optical image data closest to the dual-polarized SAR image data in terms of time, extracting the optical amphibious binary image result from the optical image data by using an optical remote sensing image vegetation and water body automatic extraction method, defining the intersection of the single-phase amphibious binary image extraction result and the optical amphibious binary image result as a combined binary image result, and performing raster vector conversion on the single-phase amphibious binary image extraction result, the optical amphibious binary image result and the combined binary image result to respectively obtain three vector images V1、V2And V3Obtaining a vector image V having an intersection relation with the vector image V31A vector diagram V' formed by all the entities in (A)1。
6. The method for amphibious SAR image boundary extraction as claimed in claim 5, further comprising, for vector diagram V3Each entity in (2) performs the following processing: let the current entity be E3Computing entity E3Has an area of A3Calculating to obtain the sum entity E3Vector diagram V' with intersection relation1Corresponding entity E in1Has an area of A1Calculating to obtain the sum entity E3Vector diagram V with intersection relation2Corresponding entity in (1)E2Has an area of A2Definition of A1And A2The smaller value therebetween is AsDefinition of A1And A2The larger value in between is AbDefinition of degree of inclusion INC ═ A3/AsThe degree of expansion is defined as EXP ═ Ab/AsIf INC<0.8 or EXP>500%, entity E1From vector diagram V1Is deleted.
7. The method for extracting the amphibious boundary of the dual-polarized SAR image according to claim 1, further comprising preprocessing the dual-polarized SAR image data after the dual-polarized SAR image data of the target region is acquired.
8. An amphibious boundary extraction system of a dual-polarized SAR image is characterized by comprising:
the data acquisition device is used for acquiring dual-polarized SAR image data of a target area;
a data analysis device for counting two types of single-polarized SAR image data contained in the dual-polarized SAR image data, respectively obtaining frequency histograms of the first type of polarized data and the second type of polarized data, and extracting a first water peak value w from the frequency histogram of the first type of polarized data1First land peak l1And a first valley value v1Extracting a second water body peak value w from the frequency histogram of the second type polarization data2Second land peak l2And a second valley value v2；
The data modeling device is used for establishing a Cartesian coordinate system by taking the first type of polarization data as an X axis and the second type of polarization data as a Y axis and marking W (W)1，w2)、L(l1，l2) And V (V)1，v2) Three feature points;
data extraction means for drawing a straight line k through the W point and the L point1Passing the V point perpendicular to the line k1Straight line k of2Falls on the straight line k2The bottom left dual polarized data points are plottedDivided into water bodies and falling on the straight line k2The upper right dual-polarized data points are divided into lands, and a single-time phase amphibian map extraction result is generated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110220619.3A CN112967308B (en) | 2021-02-26 | 2021-02-26 | Amphibious boundary extraction method and system for dual-polarized SAR image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110220619.3A CN112967308B (en) | 2021-02-26 | 2021-02-26 | Amphibious boundary extraction method and system for dual-polarized SAR image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112967308A true CN112967308A (en) | 2021-06-15 |
| CN112967308B CN112967308B (en) | 2023-09-19 |
Family
ID=76275779
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110220619.3A Active CN112967308B (en) | 2021-02-26 | 2021-02-26 | Amphibious boundary extraction method and system for dual-polarized SAR image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112967308B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN115147728A (en) * | 2022-06-24 | 2022-10-04 | 滁州学院 | Method for rapidly identifying small reservoir based on cooperation of radar data and optical data |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102175993A (en) * | 2011-01-28 | 2011-09-07 | 华中科技大学 | Radar scene matching feature reference map preparation method based on satellite SAR (synthetic aperture radar) images |
| CN102609933A (en) * | 2011-12-16 | 2012-07-25 | 电子科技大学 | Self-adaption coherent change detecting method of polarized synthetic aperture radar (SAR) images |
| US20140210661A1 (en) * | 2011-04-07 | 2014-07-31 | Mrinal Singh Balaji | Systems and methods for calibrating dual polarization radar systems |
| CN104123563A (en) * | 2014-07-17 | 2014-10-29 | 电子科技大学 | Cloude characteristic decomposition based polarimetric SAR (Synthetic Aperture Radar) image non-supervision classification method |
| CN104463166A (en) * | 2014-11-13 | 2015-03-25 | 环境保护部卫星环境应用中心 | Extraction method and system for target water body |
| CN105372631A (en) * | 2015-10-29 | 2016-03-02 | 中国科学院遥感与数字地球研究所 | Polarizing radar inversion method based on genetic-partial least square algorithm, and application of polarizing radar inversion method |
| CN107507209A (en) * | 2017-08-07 | 2017-12-22 | 西安理工大学 | The sketch map extracting method of Polarimetric SAR Image |
| CN110574071A (en) * | 2017-01-27 | 2019-12-13 | Ucl商业有限公司 | Device, method and system for aligning 3D data sets |
| CN111898250A (en) * | 2020-07-03 | 2020-11-06 | 武汉大学 | Estuary tailing branch of a river prediction method and device based on multi-source data fusion |
| CN112131946A (en) * | 2020-08-21 | 2020-12-25 | 长沙银汉空间科技有限公司 | Automatic extraction method for vegetation and water body information of optical remote sensing image |
-
2021
- 2021-02-26 CN CN202110220619.3A patent/CN112967308B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102175993A (en) * | 2011-01-28 | 2011-09-07 | 华中科技大学 | Radar scene matching feature reference map preparation method based on satellite SAR (synthetic aperture radar) images |
| US20140210661A1 (en) * | 2011-04-07 | 2014-07-31 | Mrinal Singh Balaji | Systems and methods for calibrating dual polarization radar systems |
| CN102609933A (en) * | 2011-12-16 | 2012-07-25 | 电子科技大学 | Self-adaption coherent change detecting method of polarized synthetic aperture radar (SAR) images |
| CN104123563A (en) * | 2014-07-17 | 2014-10-29 | 电子科技大学 | Cloude characteristic decomposition based polarimetric SAR (Synthetic Aperture Radar) image non-supervision classification method |
| CN104463166A (en) * | 2014-11-13 | 2015-03-25 | 环境保护部卫星环境应用中心 | Extraction method and system for target water body |
| CN105372631A (en) * | 2015-10-29 | 2016-03-02 | 中国科学院遥感与数字地球研究所 | Polarizing radar inversion method based on genetic-partial least square algorithm, and application of polarizing radar inversion method |
| CN110574071A (en) * | 2017-01-27 | 2019-12-13 | Ucl商业有限公司 | Device, method and system for aligning 3D data sets |
| CN107507209A (en) * | 2017-08-07 | 2017-12-22 | 西安理工大学 | The sketch map extracting method of Polarimetric SAR Image |
| CN111898250A (en) * | 2020-07-03 | 2020-11-06 | 武汉大学 | Estuary tailing branch of a river prediction method and device based on multi-source data fusion |
| CN112131946A (en) * | 2020-08-21 | 2020-12-25 | 长沙银汉空间科技有限公司 | Automatic extraction method for vegetation and water body information of optical remote sensing image |
Non-Patent Citations (3)
| Title |
|---|
| VINAY RAVINDRA: "A Dual-Polarization X -Band Traveling-Wave Antenna Panel for Small-Satellite Synthetic Aperture Radar", 《IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION 》 * |
| 曹须: "中国极化电子离子对撞机计划", 《核技术》 * |
| 谷鑫志: "多极化星载SAR森林覆盖变化检测方法", 《林业科学》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN114677401B (en) * | 2022-03-28 | 2022-10-21 | 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 |
| CN115147728A (en) * | 2022-06-24 | 2022-10-04 | 滁州学院 | Method for rapidly identifying small reservoir based on cooperation of radar data and optical data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112967308B (en) | 2023-09-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Sheng et al. | Quantitative dynamic flood monitoring with NOAA AVHRR | |
| Puniach et al. | Application of UAV-based orthomosaics for determination of horizontal displacement caused by underground mining | |
| CN107507200B (en) | SAR image high-precision large-range water area extraction method based on connectivity detection and noise suppression | |
| Capaldo et al. | DSM generation from high resolution imagery: applications with WorldView-1 and Geoeye-1. | |
| CN110988909B (en) | TLS-based vegetation coverage measuring method for sand vegetation in severe cold fragile area | |
| CN114387516B (en) | Single-season rice SAR (synthetic aperture radar) identification method for small and medium-sized fields in complex terrain environment | |
| CN112967308A (en) | Amphibious SAR image boundary extraction method and system | |
| Liu et al. | Detection of rapid erosion in SE Spain | |
| CN117523404B (en) | Method, device, terminal and storage medium for monitoring dynamic change of storage yard | |
| Krauss et al. | Fusing stereo and multispectral data from WorldView-2 for urban modeling | |
| Ukhnaa et al. | Modification of urban built-up area extraction method based on the thematic index-derived bands | |
| Sun et al. | Individual tree segmentation and biomass estimation based on UAV Digital aerial photograph | |
| Gu et al. | Ground point extraction using self-adaptive-grid and point to surface comparison | |
| Lian et al. | Denoising algorithm based on local distance weighted statistics for photon counting LiDAR point data | |
| Oh et al. | Extraction of digital elevation model using stereo matching with slope-adaptive patch transformation | |
| Sohn et al. | Shadow-effect correction in aerial color imagery | |
| d'Angelo et al. | Automatic generation of high quality DSM Based on IRS-P5 Cartosat-1 Stereo Data | |
| Liang et al. | Water Body Extraction Research Based on S Band SAR Satellite of | |
| Eltner | Photogrammetric techniques for across-scale soil erosion assessment: developing methods to integrate multi-temporal high resolution topography data at field plots | |
| Cheng et al. | Generation of pixel-level SAR image time series using a locally adaptive matching technique | |
| Heckel et al. | The first sub-meter resolution digital elevation model of the Kruger National Park, South Africa | |
| Dare | New techniques for the automatic registration of microwave and optical remotely sensed images | |
| Bolla et al. | The tracking and prediction of high intensity rainstorms | |
| Li et al. | A high-resolution satellite DEM filtering method assisted with building segmentation | |
| James | Three-Dimensional Reconstruction of Braided River Morphology and Morphodynamics with Structure-from-Motion Photogrammetry |
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 |