CN117392433A - Fine recognition method for different freshwater resource types by combining SAR (synthetic aperture radar) and optical image - Google Patents
Fine recognition method for different freshwater resource types by combining SAR (synthetic aperture radar) and optical image Download PDFInfo
- Publication number
- CN117392433A CN117392433A CN202311188266.9A CN202311188266A CN117392433A CN 117392433 A CN117392433 A CN 117392433A CN 202311188266 A CN202311188266 A CN 202311188266A CN 117392433 A CN117392433 A CN 117392433A
- Authority
- CN
- China
- Prior art keywords
- fresh water
- image
- sar
- flooding
- calculating
- 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
- 239000013505 freshwater Substances 0.000 title claims abstract description 157
- 230000003287 optical effect Effects 0.000 title claims abstract description 67
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000003066 decision tree Methods 0.000 claims abstract description 29
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 28
- 230000003595 spectral effect Effects 0.000 claims abstract description 21
- 230000011218 segmentation Effects 0.000 claims abstract description 12
- 238000010586 diagram Methods 0.000 claims description 22
- 238000004364 calculation method Methods 0.000 claims description 16
- 230000002194 synthesizing effect Effects 0.000 claims description 14
- 238000007781 pre-processing Methods 0.000 claims description 13
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 12
- 239000002131 composite material Substances 0.000 claims description 9
- 238000012937 correction Methods 0.000 claims description 9
- 238000004519 manufacturing process Methods 0.000 claims description 9
- 238000002310 reflectometry Methods 0.000 claims description 8
- 230000010287 polarization Effects 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 4
- 238000006424 Flood reaction Methods 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 238000012876 topography Methods 0.000 claims description 3
- 238000003860 storage Methods 0.000 claims description 2
- 238000012634 optical imaging Methods 0.000 claims 3
- 238000009826 distribution Methods 0.000 abstract description 10
- 238000001228 spectrum Methods 0.000 abstract description 5
- 230000009286 beneficial effect Effects 0.000 abstract description 4
- 238000000605 extraction Methods 0.000 description 4
- 238000003384 imaging method Methods 0.000 description 4
- 238000012544 monitoring process Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000011835 investigation Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000003306 harvesting Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/10—Image acquisition
- G06V10/12—Details of acquisition arrangements; Constructional details thereof
- G06V10/14—Optical characteristics of the device performing the acquisition or on the illumination arrangements
- G06V10/143—Sensing or illuminating at different wavelengths
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
- G06V10/803—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level of input or preprocessed data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
- G06V10/806—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/182—Network patterns, e.g. roads or rivers
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Databases & Information Systems (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Image Processing (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention relates to a method for finely identifying different freshwater resource types by combining SAR and optical images, which comprises the following steps: acquiring a time sequence SAR image and a time sequence optical image and corresponding synthesized images; constructing a normalized flooding index NDHF; performing multi-scale segmentation on the synthesized image to generate a homogeneous unit object of the image; identifying the minimum range of fresh water resources by adopting a threshold method; constructing a decision tree algorithm based on spectral characteristics to identify a flooding area; and constructing a decision tree algorithm based on geometric shape characteristics to identify rivers, lakes, ponds and culture ponds. The beneficial effects of the invention are as follows: the invention comprehensively uses SAR backscattering characteristics, optical spectrum characteristics and geometric shape characteristics to construct different decision tree distinguishing algorithms so as to obtain the careful distinction of different fresh water resource types and the spatial range of different fresh water resource type distribution, thereby being beneficial to timely and accurately grasping the distribution and change profile of the existing fresh water resource and realizing reasonable distribution and scientific scheduling of the fresh water resource.
Description
Technical Field
The invention relates to the technical field of remote sensing image processing, in particular to a method for finely identifying different freshwater resource types by combining SAR and optical images.
Background
The fresh water ecological system has the functions of regulating climate, conserving water sources, protecting fresh water biodiversity and the like, and is a foundation for stable production and harvest of agriculture and industrial development. With the increasing prominence of global water resource shortage, fine identification and management of fresh water resources has become particularly important. The freshwater ecosystem can be divided into subsystems such as lake ecosystem, pond ecosystem (pond and culture pond), river ecosystem (river and flooding area) and the like according to the water surface type.
The optical remote sensing imaging range is large, the advantages of high spatial resolution and abundant spectral information are achieved, the observation cost is low, the short-time repeated observation can be realized, and the optical remote sensing imaging method is widely applied to aspects of fresh water resource investigation and monitoring. The high spatial resolution and spectral characteristics of multispectral remote sensing have strong recognition capability on the microcosmic aspect, but multispectral remote sensing is finally influenced by weather, sunlight and the like, and all-weather imaging can not be realized all the day. The synthetic aperture radar SAR can penetrate through cloud layers to reflect the backward scattering information of ground objects, and can make up for the defect that optical remote sensing cannot observe all the time and all the weather in monitoring. In addition, the SAR satellite images are all formed by side view imaging, so that the shape and structure characteristics of the ground object can be reflected well, and the ground object information identification is facilitated. Therefore, the single-source remote sensing data can not solve the problem of fine recognition of fresh water resources.
Disclosure of Invention
The invention aims at overcoming the defects of the prior art, and provides a method for finely identifying different freshwater resource types by combining SAR and optical images.
In a first aspect, a method for finely identifying different freshwater resource types by combining SAR and optical images is provided, comprising:
step 1, acquiring a sequential SAR image and a sequential optical image, preprocessing the sequential SAR image and the sequential optical image, generating a corresponding median, maximum and minimum synthesized image for the preprocessed sequential SAR image, and generating a corresponding median synthesized image for the preprocessed sequential optical image;
step 2, synthesizing an image according to the maximum value and the minimum value of the time sequence SAR image, selecting fresh water and non-fresh water type sample points, acquiring SAR backward scattering coefficient data, and constructing a normalized flooding index NDHF;
step 3, multi-scale segmentation is carried out on the synthesized image, and a homogeneous unit object of the image is generated;
step 4, calculating spectral characteristics of the homogeneous unit object, combining gradient parameters of digital elevation model DEM data, and identifying a minimum range of fresh water resources by adopting a threshold method;
step 5, synthesizing images by relying on the maximum value and the minimum value of SAR, calculating the flooding index NDHF of each homogeneous unit object, and constructing a decision tree algorithm based on spectral characteristics to identify a flooding area;
step 6, merging homogeneous unit objects, calculating the geometric shape characteristics of each homogeneous unit object, and constructing a decision tree algorithm based on the geometric shape characteristics to identify rivers, lakes, ponds and culture ponds;
and 7, converting the identified fresh water raster data into vector data to obtain identification results of different fresh water resource types.
Preferably, in step 1, the sequential SAR image is a Sentinel-1 satellite GRD image, the band is a C band, the polarization mode is VH polarization, and the preprocessing of the sequential SAR image includes: track correction, thermal noise removal, radiometric calibration, doppler topography correction, lee filtering and decibelization; the time sequence optical image is a Sentinel-2 satellite MSI image, the wave bands are red, green, blue and near infrared, and the preprocessing of the time sequence optical image comprises the following steps: radiometric calibration, atmospheric correction, band superposition, mosaic and clipping.
Preferably, in step 2, the calculation formula of the normalized flooding index NDHF is:
wherein VH min Is the VH backscattering coefficient minimum value of the SAR minimum value synthesized image, VH max Is the VH backscatter coefficient maximum value of the SAR maximum value synthesized image.
Preferably, in step 3, the parameters of the multi-scale segmentation include: scale, shape, compactness, and the weight of the composite image.
Preferably, step 4 includes:
step 4.1, calculating the gradient by using digital elevation model DEM data, wherein a gradient calculation formula is as follows:
slope=arctan (elevation difference/horizontal distance)
Wherein Slope represents a gradient, arctan represents an arctangent function;
step 4.2, calculating a normalized water index NDWI, wherein the formula is as follows:
wherein P is green Representing the reflectivity of the green band, P nir Indicating the reflectivity of the near infrared band;
and 4.3, calculating relevant parameters of the fresh water and non-fresh water type sample points, determining an optimal threshold value, and identifying all fresh water.
Preferably, step 4.3 comprises:
step 4.3.1, calculating pixel values of VH corresponding pixels in a median synthetic image of the SAR of fresh water and non-fresh water type sample points, manufacturing a box diagram, and determining a threshold B1 by the average value of the lower quartile of the non-fresh water resource and the upper quartile of the fresh water resource;
step 4.3.2, calculating gradients of fresh water and non-fresh water type sample points, drawing a histogram, and determining a gradient threshold value as B2;
step 4.3.3, calculating pixel values of pixels corresponding to the normalized water index NDWI of the fresh water and non-fresh water type sample points in the optical median synthetic image, manufacturing a box diagram, and determining a threshold B3 by the average value of the lower quartile of the fresh water resource and the upper quartile of the non-fresh water resource;
and 4.3.4, calculating pixel values of pixels corresponding to near infrared wave bands of fresh water and non-fresh water type sample points in an optical median synthetic image, manufacturing a box diagram, determining a threshold B4 by the average value of the upper quartile of fresh water resources and the lower quartile of non-fresh water resources, and identifying all fresh water.
Preferably, in step 5, according to the decision tree algorithm model of the spectral feature, the decision tree algorithm model has a flooding index threshold B5 and an annual VH backscattering coefficient maximum threshold B6, and the determining manner of the flooding index threshold B5 is as follows:
calculating flooding indexes of a flooding area and other fresh water sample points, and making a box diagram, wherein the average value of the lower quartile of the flooding indexes of the flooding sample points and the upper quartile of the flooding indexes of other fresh water is a threshold B5 for distinguishing flooding from other fresh water resource types; the VH backscattering coefficient maximum value threshold B6 is determined in the following manner: calculating the maximum value of the VH backscattering coefficient of the sample points of the flooding area and the non-flooding area, and making a box diagram, and flooding the VH of the sample points max Upper quartile of (v) and VH of non-flooded sample points max The mean of the lower quartile of (c) is a threshold B6 that determines the flooding zone and other freshwater resource types.
Preferably, step 6 includes:
step 6.1, merging the homogeneous unit objects, and performing post-classification treatment;
step 6.2, calculating the rectangular fitting degree, the ellipse fitting degree and the area of each homogeneous unit object;
step 6.3, according to a decision tree algorithm model of geometric shape characteristics, determining a threshold value F1 of rectangular fitting degree and a threshold value F2 of ellipse fitting degree, wherein the determination mode of the threshold value F1 of rectangular fitting degree and the threshold value F2 of ellipse fitting degree is as follows:
step 6.3.1, calculating rectangular fitting degree and ellipse fitting degree of sample points of rivers, lakes, reservoirs, floods, ponds and culture ponds, and making a scatter diagram;
step 6.3.2, determining that the average value of the maximum value of the rectangular fitting degree of the river sample points and the minimum value of the rectangular fitting degree of the sample points of the lakes, the ponds and the culture ponds is a threshold value F1 of the rectangular fitting degree;
and 6.3.3, determining the average value of the maximum value of the ellipse fitting degree of the river sample points and the minimum value of the ellipse fitting degree of the sample points of the lakes, the ponds and the culture ponds as a threshold value F2 of the ellipse fitting degree.
In a second aspect, a system for finely identifying different freshwater resource types of a joint SAR and an optical image is provided, and the method for finely identifying different freshwater resource types of the joint SAR and the optical image in the first aspect is performed, and includes:
the acquisition module is used for acquiring the time sequence SAR image and the time sequence optical image, preprocessing the time sequence SAR image and the time sequence optical image, generating corresponding median, maximum and minimum synthesized images for the preprocessed time sequence SAR image, and generating corresponding median synthesized images for the preprocessed time sequence optical image;
the construction module is used for synthesizing the images according to the maximum value and the minimum value of the time sequence SAR image, acquiring SAR backward scattering coefficient data and constructing a normalized flooding index NDHF;
the generation module is used for carrying out multi-scale segmentation on the synthesized image to generate a homogeneous unit object of the image;
the first calculation module is used for calculating the spectral characteristics of the homogeneous unit object, combining gradient parameters of Digital Elevation Model (DEM) data, and identifying the minimum range of fresh water resources by adopting a threshold method;
the second calculation module is used for synthesizing images by relying on the maximum value and the minimum value of SAR, calculating the flooding index NDHF of each homogeneous unit object, and constructing a decision tree algorithm based on spectral characteristics to identify a flooding area;
the third calculation module is used for merging the homogeneous unit objects, calculating the geometric shape characteristics of each homogeneous unit object, and constructing a decision tree algorithm based on the geometric shape characteristics to identify rivers, lakes, ponds and culture ponds;
and the conversion module is used for converting the identified fresh water raster data into vector data to obtain identification results of different fresh water resource types.
In a third aspect, a computer storage medium having a computer program stored therein is provided; the computer program when run on a computer causes the computer to execute the method for finely identifying different freshwater resource types of the combined SAR and optical image according to any one of the first aspect.
The beneficial effects of the invention are as follows: based on the time sequence SAR and the optical image, the SAR backscattering characteristics, the optical spectrum characteristics and the geometric shape characteristics of different fresh water resources are extracted, and the decision tree algorithm for distinguishing different fresh water resource types is constructed by comprehensively using the SAR backscattering characteristics, the optical spectrum characteristics and the geometric shape characteristics, so that the fine distinguishing of different fresh water resource types and the spatial range of different fresh water resource type distribution are obtained, the current fresh water resource distribution and change profile can be mastered timely and accurately, and reasonable distribution and scientific scheduling of fresh water resources are realized. The method has a certain reference significance for the fine and automatic extraction of the fresh water in a large-scale or terrain complex area.
Drawings
FIG. 1 is a flow chart of a method for fine recognition of different freshwater resource types by combining SAR and optical images;
FIG. 2 is a graph of a local effect of multi-scale segmentation of a joint SAR and optical image;
FIG. 3 is a graph showing normalized water index (NDWI), near infrared band (NIR) reflectivity, VH backscattering coefficient median values for sample points of freshwater resource types and non-freshwater resource types;
FIG. 4 is a diagram of normalized flooding index (NDHF) of a flooding domain and other fresh water resources;
FIG. 5 is a schematic diagram of a decision tree model;
FIG. 6 is a diagram showing the result of fresh water identification;
fig. 7 is a schematic diagram of extraction of different types of fresh water resources.
Detailed Description
The invention is further described below with reference to examples. The following examples are presented only to aid in the understanding of the invention. It should be noted that it will be apparent to those skilled in the art that modifications can be made to the present invention without departing from the principles of the invention, and such modifications and adaptations are intended to be within the scope of the invention as defined in the following claims.
Example 1:
complementary information of different data sources in the same area can be enriched by the multi-source remote sensing image (multispectral and SAR), so that uncertainty of a single data source is reduced, ambiguity is reduced, and accordingly complete and consistent information description of a target is formed. In addition, unlike conventional freshwater resource identification and extraction, the essence of fine classification requires that the remote sensing technology not only realize the fine distinction of different freshwater resource types, but also can accurately judge the spatial range of the distribution of different freshwater resource types. The multisource remote sensing image has the information expression capability of providing more complete fresh water resources, and has the advantage of more information compared with the single-source remote sensing image to realize the fine monitoring of the fresh water resources.
The embodiment of the application provides a method for finely identifying different freshwater resource types by combining SAR and optical images, which is shown in figure 1 and comprises the following steps:
step 1, acquiring a time sequence SAR image and a time sequence optical image, preprocessing the time sequence SAR image and the time sequence optical image, generating corresponding median, maximum and minimum synthesized images for the preprocessed time sequence SAR image, and generating corresponding median synthesized images for the preprocessed time sequence optical image.
Specifically, the sequential SAR image is a Sentinel-1 satellite GRD image, the wave band is a C wave band, the polarization mode is VH polarization, and the preprocessing of the sequential SAR image comprises the following steps: track correction, thermal noise removal, radiometric calibration, doppler topography correction, lee filtering and decibelization; the time sequence optical image is a Sentinel-2 satellite MSI image, the wave bands are red, green, blue and near infrared, and the preprocessing of the time sequence optical image comprises the following steps: radiometric calibration, atmospheric correction, band superposition, mosaic and clipping.
And 2, synthesizing images according to the maximum value and the minimum value of the time sequence SAR image, selecting fresh water and non-fresh water type sample points, acquiring SAR backward scattering coefficient data, and constructing a normalized flooding index NDHF.
Specifically, depending on minimum and maximum SAR images, selecting flood areas, rivers, lakes, ponds and culture pond sample points according to google earth high-resolution images and actual investigation, acquiring SAR backward scattering coefficient data, and constructing a normalized flooding index NDHF, wherein the NDHF calculation formula is as follows:
wherein VH min Is the VH backscattering coefficient minimum value of the SAR minimum value synthesized image, VH max Is the VH backscatter coefficient maximum value of the SAR maximum value synthesized image.
And 3, performing multi-scale segmentation on the composite image to generate a homogeneous unit object of the image.
Specifically, the parameters of the multi-scale segmentation include: scale, shape, compactness, and weight of the composite image. For example, the parameter scale, shape, compactness are 5, 0.1, 0.8, respectively, and the median composite image, maximum composite image, minimum composite image, and optical median composite image weights are 1, respectively.
And 4, calculating spectral characteristics of the homogeneous unit object, combining gradient parameters of digital elevation model DEM data, and identifying the minimum range of the fresh water resource by adopting a threshold method.
And 5, synthesizing images by relying on the maximum value and the minimum value of SAR, calculating the flooding index NDHF of each homogeneous unit object, and constructing a decision tree algorithm based on spectral characteristics to identify a flooding area.
And 6, merging the homogeneous unit objects, calculating the geometric shape characteristics of each homogeneous unit object, and constructing a decision tree algorithm based on the geometric shape characteristics to identify rivers, lakes, ponds and culture ponds.
And 7, converting the identified fresh water raster data into vector data to obtain identification results of different fresh water resource types.
Example 2:
based on embodiment 1, embodiment 2 of the present application provides a more specific method for finely identifying different freshwater resource types by combining SAR and optical images, which includes:
step 1, acquiring a time sequence SAR image and a time sequence optical image, preprocessing the time sequence SAR image and the time sequence optical image, generating corresponding median, maximum and minimum synthesized images for the preprocessed time sequence SAR image, and generating corresponding median synthesized images for the preprocessed time sequence optical image.
And 2, synthesizing images according to the maximum value and the minimum value of the time sequence SAR image, selecting fresh water and non-fresh water type sample points, acquiring SAR backward scattering coefficient data, and constructing a normalized flooding index NDHF.
Step 3, as shown in fig. 2, the composite image is subjected to multi-scale segmentation to generate a homogeneous unit object of the image.
And 4, calculating spectral characteristics of the homogeneous unit object, combining gradient parameters of digital elevation model DEM data, and identifying the minimum range of the fresh water resource by adopting a threshold method.
Specifically, step 4 includes:
step 4.1, calculating the gradient by using Digital Elevation Model (DEM) data, wherein the Digital Elevation Model (DEM) data are ALOS-12.5 m DEM data, and a gradient calculation formula is as follows:
slope=arctan (elevation difference/horizontal distance)
Wherein Slope represents a gradient, arctan represents an arctangent function;
step 4.2, spectral features including VH backscattering coefficient, near infrared band (NIR) reflectivity and normalized water index (Normalized Difference Water Index, NDWI), the calculation formula of the normalized water index NDWI is:
wherein P is green Representing the reflectivity of the green band, P nir Indicating the reflectivity of the near infrared band;
and 4.3, selecting fresh water and non-fresh water type sample points, calculating related parameters, determining an optimal threshold value, and identifying all fresh water.
Step 4.3 comprises:
step 4.3.1, selecting fresh water and non-fresh water type sample points, calculating pixel values of VH corresponding pixels in a median synthetic image of SAR of the fresh water and non-fresh water type sample points, manufacturing a box diagram, and determining a threshold B1 by the average value of the lower quartile of non-fresh water resources and the upper quartile of fresh water resources;
step 4.3.2, calculating gradients of fresh water and non-fresh water type sample points, drawing a histogram, and determining a gradient threshold value as B2;
step 4.3.3, calculating pixel values of pixels corresponding to the normalized water index NDWI of the fresh water and non-fresh water type sample points in the optical median synthetic image, manufacturing a box diagram, and determining a threshold B3 by the average value of the lower quartile of the fresh water resource and the upper quartile of the non-fresh water resource;
and 4.3.4, calculating pixel values of pixels corresponding to near infrared wave bands of fresh water and non-fresh water type sample points in an optical median synthetic image, manufacturing a box diagram, determining a threshold B4 by the average value of the upper quartile of fresh water resources and the lower quartile of non-fresh water resources, and identifying all fresh water.
And 5, as shown in fig. 4, synthesizing an image by relying on the maximum value and the minimum value of SAR, calculating the flooding index NDHF of each homogeneous unit object, and constructing a decision tree algorithm of spectral characteristics to identify a flooding area.
In step 5, according to the decision tree algorithm model of the spectral feature, the method has a flooding index threshold B5 and an annual VH backscattering coefficient maximum threshold B6, and the determining manner of the flooding index threshold B5 is as follows:
calculating flooding indexes of a flooding area and other fresh water sample points, and making a box diagram, wherein the average value of the lower quartile of the flooding indexes of the flooding sample points and the upper quartile of the flooding indexes of other fresh water is an optimal threshold B5 for distinguishing flooding from other fresh water resource types;
the VH backscattering coefficient maximum value threshold B6 is determined in the following manner:
calculating the maximum value of the VH backscattering coefficient of the sample points of the flooding area and the non-flooding area, and making a box diagram, and flooding the VH of the sample points max Upper quartile of (v) and VH of non-flooded sample points max Lower quartile of (2)The mean value of (2) is a threshold B6 that determines the flooding zone and other freshwater resource types.
And 6, combining the homogeneous unit objects, calculating the geometric shape characteristics of each homogeneous unit object, and constructing a decision tree algorithm based on the geometric shape characteristics to identify rivers, lakes, ponds and culture ponds as shown in fig. 5.
The step 6 comprises the following steps:
and 6.1, merging the homogeneous unit objects, and performing post-classification treatment.
And 6.2, calculating the rectangular fitting degree, the ellipse fitting degree and the area of each homogeneous unit object.
The rectangular fitting feature and the ellipse fitting feature are specifically defined as follows:
rectangular fitting characteristics: the degree of matching of image objects to a cuboid of similar size and scale is described. 0 indicates no match and 1 indicates a perfect match. Is based on a cuboid of the same volume as the image object under consideration. The ratio of the rectangular parallelepiped is equal to the ratio of the length, width, and thickness of the image object. The volume of the image object outside the rectangle is compared with the volume inside the cuboid without the image object filled.
Ellipse fitting features describe how well an image object matches an ellipsoid of similar size and scale. 0 indicates no match and 1 indicates a perfect match. Is based on ellipsoids having the same volume as the image object under consideration. The ratio of ellipsoids is equal to the ratio of the length, width and thickness of an image object. The volume of the image objects outside the ellipsoid is compared to the volume of the image objects inside the ellipsoid that are not filled.
Step 6.3, as shown in fig. 6, according to the decision tree algorithm model of the geometric feature, the determination modes of the threshold value F1 of the rectangular fitting degree and the threshold value F2 of the elliptical fitting degree, and the determination modes of the threshold value F1 of the rectangular fitting degree and the threshold value F2 of the elliptical fitting degree are as follows:
step 6.3.1, calculating rectangular fitting degree and ellipse fitting degree of sample points of rivers, lakes, reservoirs, floods, ponds and culture ponds, and making a scatter diagram;
step 6.3.2, determining that the average value of the maximum value of the rectangular fitting degree of the river sample points and the minimum value of the rectangular fitting degree of the sample points of the lakes, the ponds and the culture ponds is a threshold value F1 of the rectangular fitting degree;
and 6.3.3, determining the average value of the maximum value of the ellipse fitting degree of the river sample points and the minimum value of the ellipse fitting degree of the sample points of the lakes, the ponds and the culture ponds as a threshold value F2 of the ellipse fitting degree.
The decision tree algorithm model S1 can take a fixed value of 1.1 hectare and S2 can take a fixed value of 8 hectare.
And 7, converting the identified fresh water raster data into vector data to obtain identification results of different fresh water resource types.
Based on the decision tree algorithm model, as shown in fig. 7, (a) is a recognition result of the type of the freshwater resource in Fuling area of Chongqing city, (B) (C) (D) (E) corresponds to the amplifying effect of the B, C, D and E in (a), and (B ') (C') (D ') (E') corresponds to the SAR image of the B, C, D and E in (a).
In this embodiment, the same or similar parts as those in embodiment 1 may be referred to each other, and will not be described in detail in this application.
Example 3:
based on embodiments 1 and 2, embodiment 3 of the present application provides a fine recognition system for different freshwater resource types combining SAR and optical images, including:
the acquisition module is used for acquiring the time sequence SAR image and the time sequence optical image, preprocessing the time sequence SAR image and the time sequence optical image, generating corresponding median, maximum and minimum synthesized images for the preprocessed time sequence SAR image, and generating corresponding median synthesized images for the preprocessed time sequence optical image;
the construction module is used for synthesizing the images according to the maximum value and the minimum value of the time sequence SAR image, acquiring SAR backward scattering coefficient data and constructing a normalized flooding index NDHF;
the generation module is used for carrying out multi-scale segmentation on the synthesized image to generate a homogeneous unit object of the image;
the first calculation module is used for calculating the spectral characteristics of the homogeneous unit object, combining gradient parameters of Digital Elevation Model (DEM) data, and identifying the minimum range of fresh water resources by adopting a threshold method;
the second calculation module is used for synthesizing images by relying on the maximum value and the minimum value of SAR, calculating the flooding index NDHF of each homogeneous unit object, and constructing a decision tree algorithm based on spectral characteristics to identify a flooding area;
the third calculation module is used for merging the homogeneous unit objects, calculating the geometric shape characteristics of each homogeneous unit object, and constructing a decision tree algorithm based on the geometric shape characteristics to identify rivers, lakes, ponds and culture ponds;
and the conversion module is used for converting the identified fresh water raster data into vector data to obtain identification results of different fresh water resource types.
Specifically, the system provided in this embodiment is a system corresponding to the methods provided in embodiments 1 and 2, and therefore, the portions in this embodiment that are the same as or similar to those in embodiment 1 may be referred to each other, and will not be described in detail in this application.
In summary, the method extracts the SAR backscattering characteristics, the optical spectrum characteristics and the geometric shape characteristics of different fresh water resources based on the time sequence SAR and the optical images, and comprehensively uses the SAR backscattering characteristics, the optical spectrum characteristics and the geometric shape characteristics to construct a decision tree algorithm for distinguishing different fresh water resource types, so as to obtain the careful distinction of different fresh water resource types and the spatial range of the distribution of different fresh water resource types, thereby being beneficial to timely and accurately grasping the distribution and the change profile of the existing fresh water resources and realizing reasonable distribution and scientific scheduling of the fresh water resources. The method has a certain reference significance for the fine and automatic extraction of the fresh water in a large-scale or terrain complex area.
Claims (10)
1. The method for finely identifying different freshwater resource types by combining SAR and optical images is characterized by comprising the following steps of:
step 1, acquiring a sequential SAR image and a sequential optical image, preprocessing the sequential SAR image and the sequential optical image, generating a corresponding median, maximum and minimum synthesized image for the preprocessed sequential SAR image, and generating a corresponding median synthesized image for the preprocessed sequential optical image;
step 2, synthesizing an image according to the maximum value and the minimum value of the time sequence SAR image, selecting fresh water and non-fresh water type sample points, acquiring SAR backward scattering coefficient data, and constructing a normalized flooding index NDHF;
step 3, multi-scale segmentation is carried out on the synthesized image, and a homogeneous unit object of the image is generated;
step 4, calculating spectral characteristics of the homogeneous unit object, combining gradient parameters of digital elevation model DEM data, and identifying a minimum range of fresh water resources by adopting a threshold method;
step 5, synthesizing images by relying on the maximum value and the minimum value of SAR, calculating the flooding index NDHF of each homogeneous unit object, and constructing a decision tree algorithm based on spectral characteristics to identify a flooding area;
step 6, merging homogeneous unit objects, calculating the geometric shape characteristics of each homogeneous unit object, and constructing a decision tree algorithm based on the geometric shape characteristics to identify rivers, lakes, ponds and culture ponds;
and 7, converting the identified fresh water raster data into vector data to obtain identification results of different fresh water resource types.
2. The method for finely identifying different freshwater resource types by combining SAR and optical images according to claim 1, wherein in the step 1, the sequential SAR image is a Sentinel-1 satellite GRD image, the wave band is a C wave band, the polarization mode is VH polarization, and the preprocessing of the sequential SAR image comprises: track correction, thermal noise removal, radiometric calibration, doppler topography correction, lee filtering and decibelization; the time sequence optical image is a Sentinel-2 satellite MSI image, the wave bands are red, green, blue and near infrared, and the preprocessing of the time sequence optical image comprises the following steps: radiometric calibration, atmospheric correction, band superposition, mosaic and clipping.
3. The method for finely identifying different freshwater resource types by combining SAR and optical images according to claim 2, wherein in step 2, the calculation formula of the normalized flooding index NDHF is:
wherein VH min Is the VH backscattering coefficient minimum value of the SAR minimum value synthesized image, VH max Is the VH backscatter coefficient maximum value of the SAR maximum value synthesized image.
4. The method for fine recognition of different freshwater resource types by combining SAR and optical imaging according to claim 3, wherein in step 3, the parameters of the multi-scale segmentation include: scale, shape, compactness, and weight of the composite image.
5. The method for fine recognition of different freshwater resource types by combining SAR and optical images according to claim 4, wherein step 4 comprises:
step 4.1, calculating the gradient by using digital elevation model DEM data, wherein a gradient calculation formula is as follows:
slope=arctan (elevation difference/horizontal distance)
Wherein Slope represents a gradient, arctan represents an arctangent function;
step 4.2, calculating a normalized water index NDWI, wherein the formula is as follows:
wherein P is green Representing the reflectivity of the green band, P nir Indicating near infrared reflectivity;
and 4.3, calculating relevant parameters of the fresh water and non-fresh water type sample points, determining an optimal threshold value, and identifying all fresh water.
6. The method for fine recognition of different freshwater resource types by combining SAR and optical imaging according to claim 5, wherein step 4.3 comprises:
step 4.3.1, calculating pixel values of VH corresponding pixels in the SAR median synthetic image of the freshwater and non-freshwater type sample points, and making a box diagram, and determining a threshold B1 by the average value of the lower quartile of the non-freshwater resource and the upper quartile of the freshwater resource;
step 4.3.2, calculating gradients of fresh water and non-fresh water type sample points, drawing a histogram, and determining a gradient threshold value as B2;
step 4.3.3, calculating pixel values of pixels corresponding to the normalized water index NDWI of the fresh water and non-fresh water type sample points in the optical median synthetic image, manufacturing a box diagram, and determining a threshold B3 by the average value of the lower quartile of the fresh water resource and the upper quartile of the non-fresh water resource;
and 4.3.4, calculating pixel values of pixels corresponding to near infrared wave bands of fresh water and non-fresh water type sample points in an optical median synthetic image, manufacturing a box diagram, determining a threshold B4 by the average value of the upper quartile of fresh water resources and the lower quartile of non-fresh water resources, and identifying all fresh water.
7. The method for finely identifying different freshwater resource types by combining SAR and optical images according to claim 6, wherein in step 5, according to the decision tree algorithm model of spectral characteristics, the method comprises a flooding index threshold B5 and an annual VH backscattering coefficient maximum threshold B6, and the method for determining the flooding index threshold B5 is as follows:
calculating flooding indexes of a flooding area and other fresh water sample points, and making a box diagram, wherein the average value of the lower quartile of the flooding indexes of the flooding sample points and the upper quartile of the flooding indexes of other fresh water is a threshold B5 for distinguishing flooding from other fresh water resource types;
the annual VH backscatter coefficient maximum value threshold B6 is determined in the following manner: calculating the maximum value of the VH backscattering coefficient of the sample points of the flooding area and the non-flooding area, and making a box diagram, and flooding the VH of the sample points max Upper quartile of (v) and VH of non-flooded sample points max The mean of the lower quartile of (c) is a threshold B6 that determines flooding and other freshwater resource types.
8. The method for fine recognition of different freshwater resource types by combining SAR and optical images according to claim 7, wherein step 6 comprises:
step 6.1, merging the homogeneous unit objects, and performing post-classification treatment;
step 6.2, calculating the rectangular fitting degree, the ellipse fitting degree and the area of each homogeneous unit object;
step 6.3, according to a decision tree algorithm model of geometric shape characteristics, determining a threshold value F1 of rectangular fitting degree and a threshold value F2 of ellipse fitting degree, wherein the determination mode of the threshold value F1 of rectangular fitting degree and the threshold value F2 of ellipse fitting degree is as follows:
step 6.3.1, calculating rectangular fitting degree and ellipse fitting degree of sample points of rivers, lakes, reservoirs, floods, ponds and culture ponds, and making a scatter diagram;
step 6.3.2, determining that the average value of the maximum value of the rectangular fitting degree of the river sample points and the minimum value of the rectangular fitting degree of the sample points of the lakes, the ponds and the culture ponds is a threshold value F1 of the rectangular fitting degree;
and 6.3.3, determining the average value of the maximum value of the ellipse fitting degree of the river sample points and the minimum value of the ellipse fitting degree of the sample points of the lakes, the ponds and the culture ponds as a threshold value F2 of the ellipse fitting degree.
9. A system for finely identifying different freshwater resource types of a joint SAR and an optical image, for performing the finely identifying method of different freshwater resource types of a joint SAR and an optical image according to any one of claims 1 to 8, comprising:
the acquisition module is used for acquiring the time sequence SAR image and the time sequence optical image, preprocessing the time sequence SAR image and the time sequence optical image, generating corresponding median, maximum and minimum synthesized images for the preprocessed time sequence SAR image, and generating corresponding median synthesized images for the preprocessed time sequence optical image;
the construction module is used for synthesizing the images according to the maximum value and the minimum value of the time sequence SAR image, acquiring SAR backward scattering coefficient data and constructing a normalized flooding index NDHF;
the generation module is used for carrying out multi-scale segmentation on the synthesized image to generate a homogeneous unit object of the image;
the first calculation module is used for calculating the spectral characteristics of the homogeneous unit object, combining gradient parameters of Digital Elevation Model (DEM) data, and identifying the minimum range of fresh water resources by adopting a threshold method;
the second calculation module is used for synthesizing images by relying on the maximum value and the minimum value of SAR, calculating the flooding index NDHF of each homogeneous unit object, and constructing a decision tree algorithm based on spectral characteristics to identify a flooding area;
the third calculation module is used for merging the homogeneous unit objects, calculating the geometric shape characteristics of each homogeneous unit object, and constructing a decision tree algorithm based on the geometric shape characteristics to identify rivers, lakes, ponds and culture ponds;
and the conversion module is used for converting the identified fresh water raster data into vector data to obtain identification results of different fresh water resource types.
10. A computer storage medium having a computer program stored therein; the computer program, when run on a computer, causes the computer to perform the method for finely identifying different freshwater resource types for joint SAR and optical imaging according to any of claims 1 to 8.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311188266.9A CN117392433B (en) | 2023-09-15 | 2023-09-15 | Fine recognition method for different freshwater resource types by combining SAR (synthetic aperture radar) and optical image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311188266.9A CN117392433B (en) | 2023-09-15 | 2023-09-15 | Fine recognition method for different freshwater resource types by combining SAR (synthetic aperture radar) and optical image |
Publications (2)
Publication Number | Publication Date |
---|---|
CN117392433A true CN117392433A (en) | 2024-01-12 |
CN117392433B CN117392433B (en) | 2024-06-21 |
Family
ID=89438039
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202311188266.9A Active CN117392433B (en) | 2023-09-15 | 2023-09-15 | Fine recognition method for different freshwater resource types by combining SAR (synthetic aperture radar) and optical image |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117392433B (en) |
Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109472304A (en) * | 2018-10-30 | 2019-03-15 | 厦门理工学院 | Tree species classification method, device and equipment based on SAR Yu optical remote sensing time series data |
CN109613513A (en) * | 2018-12-20 | 2019-04-12 | 长安大学 | A kind of potential landslide automatic identifying method of optical remote sensing for taking InSAR deformation into account |
KR102109917B1 (en) * | 2018-11-16 | 2020-05-12 | 대한민국(관리부서 : 환경부 국립환경과학원장) | Method for automatic inland water boundary extraction using remote sensed hyperspectral images |
CN114627380A (en) * | 2022-04-02 | 2022-06-14 | 杭州电子科技大学 | Rice identification method based on fusion of optical image and SAR time sequence data |
CN114724049A (en) * | 2022-04-11 | 2022-07-08 | 中国科学院南京地理与湖泊研究所 | Inland culture pond water surface identification method based on high-resolution remote sensing image data |
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 |
CN115223059A (en) * | 2022-08-31 | 2022-10-21 | 自然资源部第三航测遥感院 | Multi-cloud-fog-area crop planting mode extraction method based on multi-element remote sensing image |
CN115457401A (en) * | 2022-08-30 | 2022-12-09 | 宁波大学 | SAR remote sensing fine identification method for different fresh water resource types |
KR102496740B1 (en) * | 2022-06-02 | 2023-02-07 | 대한민국 | System and method for reservoir water body analysis using synthetic aperture radar data |
KR102540762B1 (en) * | 2022-10-14 | 2023-06-14 | 대한민국 | Reservoir monitoring method using satellite informations |
CN116704369A (en) * | 2023-06-07 | 2023-09-05 | 河南城建学院 | Object-oriented optical and SAR remote sensing image fusion flood extraction method and system |
-
2023
- 2023-09-15 CN CN202311188266.9A patent/CN117392433B/en active Active
Patent Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109472304A (en) * | 2018-10-30 | 2019-03-15 | 厦门理工学院 | Tree species classification method, device and equipment based on SAR Yu optical remote sensing time series data |
KR102109917B1 (en) * | 2018-11-16 | 2020-05-12 | 대한민국(관리부서 : 환경부 국립환경과학원장) | Method for automatic inland water boundary extraction using remote sensed hyperspectral images |
CN109613513A (en) * | 2018-12-20 | 2019-04-12 | 长安大学 | A kind of potential landslide automatic identifying method of optical remote sensing for taking InSAR deformation into account |
CN114627380A (en) * | 2022-04-02 | 2022-06-14 | 杭州电子科技大学 | Rice identification method based on fusion of optical image and SAR time sequence data |
CN114724049A (en) * | 2022-04-11 | 2022-07-08 | 中国科学院南京地理与湖泊研究所 | Inland culture pond water surface identification method based on high-resolution remote sensing image data |
KR102496740B1 (en) * | 2022-06-02 | 2023-02-07 | 대한민국 | System and method for reservoir water body analysis using synthetic aperture radar data |
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 |
CN115457401A (en) * | 2022-08-30 | 2022-12-09 | 宁波大学 | SAR remote sensing fine identification method for different fresh water resource types |
CN115223059A (en) * | 2022-08-31 | 2022-10-21 | 自然资源部第三航测遥感院 | Multi-cloud-fog-area crop planting mode extraction method based on multi-element remote sensing image |
KR102540762B1 (en) * | 2022-10-14 | 2023-06-14 | 대한민국 | Reservoir monitoring method using satellite informations |
CN116704369A (en) * | 2023-06-07 | 2023-09-05 | 河南城建学院 | Object-oriented optical and SAR remote sensing image fusion flood extraction method and system |
Non-Patent Citations (4)
Title |
---|
ANDREW WHYTE: "A new synergistic approach for monitoring wetlands using Sentinels -1 and 2 data with object-based machine learning algorithms", ENVIRONMENTAL MODELLING & SOFTWARE, 22 March 2018 (2018-03-22) * |
LIHUA WANG: "Detailed Processes of Tidal Flat Geomorphology Evolution Based on Time-Series Satellite Images", 《REMOTE SENSING APPLICATION IN COASTAL GEOMORPHOLOGY AND PROCESSES》, 1 September 2022 (2022-09-01) * |
李颖: "基于 Sentinel-1 数据的典型地物特征分析与洪水区面积提取", 气象与环境科学, 30 November 2022 (2022-11-30), pages 4 * |
陶圆: "结合光学与雷达遥感数据的覆膜农田机器学习分类制图对比研究", 中国优秀硕士学位论文全文数据库, 15 March 2022 (2022-03-15), pages 13 * |
Also Published As
Publication number | Publication date |
---|---|
CN117392433B (en) | 2024-06-21 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Kulkarni et al. | Pixel level fusion techniques for SAR and optical images: A review | |
Karvonen | Baltic sea ice concentration estimation using SENTINEL-1 SAR and AMSR2 microwave radiometer data | |
Fearns et al. | Shallow water substrate mapping using hyperspectral remote sensing | |
Rodrigues et al. | Use of multi-sensor data to identify and map tropical coastal wetlands in the Amazon of Northern Brazil | |
Tsokas et al. | SAR data applications in earth observation: An overview | |
CN106815559B (en) | Method and device for monitoring oyster arrangement area by utilizing SAR data, and user equipment | |
CN107688776B (en) | Urban water body extraction method | |
Demir et al. | Integrated shoreline extraction approach with use of Rasat MS and SENTINEL-1A SAR Images | |
Moniruzzaman et al. | Imaging and classification techniques for seagrass mapping and monitoring: A comprehensive survey | |
CN114998365A (en) | Ground feature classification method based on polarimetric interference SAR | |
Zhang et al. | A multiband model with successive projections algorithm for bathymetry estimation based on remotely sensed hyperspectral data in Qinghai Lake | |
Chen et al. | Decision tree-based classification in coastal area integrating polarimetric SAR and optical data | |
CN117392433B (en) | Fine recognition method for different freshwater resource types by combining SAR (synthetic aperture radar) and optical image | |
CN115457401B (en) | SAR remote sensing fine recognition method for different fresh water resource types | |
Haris et al. | COMPARISON OF SENTINEL-2 AND MULTITEMPORAL SENTINEL-1 SAR IMAGERY FOR MAPPING AQUACULTURE POND DISTRIBUTION IN THE COASTAL REGION OF BREBES REGENCY, CENTRAL JAVA, INDONESIA. | |
Vinayaraj et al. | Development of an automatic dynamic global water mask using landsat-8 images | |
Zakhvatkina et al. | SAR Sea Ice Type Classification and Drift Retrieval in the Arctic | |
Mallet et al. | Airborne LiDAR data processing | |
Saied et al. | Digital Elevation Model Enhancement using CNN-Based Despeckled SAR Images | |
Hong et al. | Mangrove extraction from super-resolution images generated by deep learning models | |
CN117173584B (en) | Land small micro water body extraction method and device for fusion of PolSAR and Pan images | |
Nieto et al. | Classifying benthic habitats and deriving bathymetry at the Caribbean Netherlands using multispectral imagery. Case study of St. Eustatius | |
Hahmann et al. | Automatic extraction of water bodies from TerraSAR-X data | |
Singh et al. | Backscatter and coherence analysis using space borne C-band data for forest characterization | |
CN113050090B (en) | Dual-polarized HH, HV radar image feature fusion enhancement method |
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 |