CN112270301A - Landsat remote sensing image water body extraction method based on wave band characteristics - Google Patents

Landsat remote sensing image water body extraction method based on wave band characteristics Download PDF

Info

Publication number
CN112270301A
CN112270301A CN202011284176.6A CN202011284176A CN112270301A CN 112270301 A CN112270301 A CN 112270301A CN 202011284176 A CN202011284176 A CN 202011284176A CN 112270301 A CN112270301 A CN 112270301A
Authority
CN
China
Prior art keywords
water body
wave bands
remote sensing
water
landsat
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202011284176.6A
Other languages
Chinese (zh)
Other versions
CN112270301B (en
Inventor
陈君
张丽丽
魏雅雪
王高旭
洪勇豪
廖洋
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hohai University HHU
Original Assignee
Hohai University HHU
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Hohai University HHU filed Critical Hohai University HHU
Priority to CN202011284176.6A priority Critical patent/CN112270301B/en
Publication of CN112270301A publication Critical patent/CN112270301A/en
Application granted granted Critical
Publication of CN112270301B publication Critical patent/CN112270301B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G06V20/182—Network patterns, e.g. roads or rivers
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/136—Segmentation; Edge detection involving thresholding
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G06V20/176—Urban or other man-made structures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G06V20/188—Vegetation
    • 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
    • G06T2207/10—Image acquisition modality
    • G06T2207/10032—Satellite or aerial image; Remote sensing
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Probability & Statistics with Applications (AREA)
  • Image Processing (AREA)
  • Geophysics And Detection Of Objects (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a method for extracting Landsat remote sensing image water body based on wave band characteristics, which comprises the following steps: selecting six common wave bands in Landsat7ETM + and Landsat 8OLI remote sensing images, analyzing the spectral sensitivity of two types of typical water bodies, determining the selection of the wave bands and giving initial weight values to the selected wave bands; performing spectrum sensitivity analysis on five typical objects except water to determine common sensitive wave bands and initial weights corresponding to the wave bands; determining a water body extraction model based on a multi-objective optimization algorithm idea and on the principle of better distinguishing water bodies from other ground objects; and finally, performing binary segmentation on the water body through the fused spectral information and the 0 threshold value to realize water body extraction. According to the invention, the construction of the water body extraction model is carried out based on the spectrum statistical characteristics of actual remote sensing image data, so that the effective distinguishing of different ground objects is achieved, the water body extraction precision is improved, and the trouble that the traditional water body extraction model is difficult to determine the optimal threshold value is avoided.

Description

Landsat remote sensing image water body extraction method based on wave band characteristics
Technical Field
The invention relates to a method for extracting a Landsat remote sensing image water body based on wave band characteristics, and belongs to the field of remote sensing image target extraction.
Background
Lakes, as an important component of the land ecosystem, play an important role in the water circulation in the natural world. The lakes and the basins thereof not only can adjust the change of the ecological environment, but also are important natural resources on which human beings live and develop socially. The spatial and temporal distribution and morphological changes of lakes are important information carriers for revealing the climatic conditions of the watershed and human activities. The lake form can be mastered in time, which is of great significance to human production and life and sustainable development of ecological environment. Before the development of remote sensing technology, researchers generally acquire morphological information of lakes through field exploration, and the method is difficult to meet the requirements on efficiency and precision. The gradual development and popularization of the remote sensing technology enable researchers to realize the superiority of the remote sensing technology, and the remote sensing technology is used for realizing the water body extraction of lakes. Compared with the traditional monitoring method, the lake monitored by the remote sensing technology has the advantages of universality, timeliness and objectivity.
The Landsat remote sensing images are rich in Landsat7 satellites and Landsat 8 satellites, and Landsat 8 carries two sensors OLI and TIRS. The OLI sensor comprises nine wave bands, and the spatial resolution of all the wave bands except the panchromatic wave band is 15m, and the spatial resolution of all the wave bands is 30 m. The Landsat7 and Landsat 8 have short return visit period and wide application, so the method can be used as a data source for extracting the lake water body. The existing lake water body extraction model comprises a single-band threshold value method, a normalized water body index method, a novel water body index method and the like, and the extraction methods have the phenomena of water body extraction omission and non-water body extraction error at different degrees.
Disclosure of Invention
The purpose of the invention is as follows: aiming at the defects of the existing water body index model, the invention aims to provide a method for extracting the Landsat remote sensing image water body based on the wave band characteristics, so as to solve the problems of extraction omission, wrong extraction, difficulty in determining an optimal threshold value and the like when the existing water body index model extracts the water body; by utilizing the abundant wave band information of the Landsat remote sensing image, the problem of false extraction in the water body extraction process is reduced, and the water body extraction precision of the Landsat remote sensing image is further improved.
The technical scheme is as follows: in order to achieve the purpose, the Landsat remote sensing image water body extraction method based on the wave band characteristics comprises the following steps:
(1) selecting six common wave bands in Landsat7ETM + and Landsat 8OLI remote sensing images, carrying out spectrum sensitivity analysis on two types of water bodies of a lake center and a shallow water area one by one on the common wave bands, determining the selection of the sensitive wave bands of each type of object by utilizing the spectrum information difference of the same object on different wave bands and giving an initial weight to the selected wave bands;
(2) spectral sensitivity analysis of five typical types of ground objects, namely high-reflection buildings, low-reflection buildings, vegetation, bare land and shadow, except a water body is carried out on the common wave bands of Landsat7ETM + and Landsat 8OLI remote sensing images, and the common sensitive wave bands of the five types of ground objects and initial weights corresponding to the wave bands are determined by combining the sensitive wave bands of the water body;
(3) based on the idea of a multi-objective optimization algorithm, a principle of better distinguishing two types of water bodies and five types of typical ground objects is used, sensitive wave bands of the water bodies and common sensitive wave bands of the five types of typical ground objects are combined, weights of the wave bands are determined, and a water body extraction model is constructed;
(4) calculating new spectral information of each pixel point according to the water body extraction model, thereby obtaining a multiband fused remote sensing image based on the water body extraction model;
(5) and according to the difference of the calculated values of the water body extraction model, taking 0 as a threshold value, distinguishing the water body and the non-water body by threshold value segmentation, taking the water body with the spectral information larger than 0 and the non-water body with the spectral information smaller than 0, and generating a binary image according to a binary segmentation result to realize water body extraction.
Further, the step (1) includes:
(11) selecting six common wave bands B2-B7 of Landsat7ETM + and Landsat 8OLI remote sensing images, respectively corresponding to Blue, Green, Red, NIR, SWIR-1 and SWIR-2 wave bands, and determining specific water body types and typical object types for spectral sensitivity analysis;
(12) and (3) carrying out spectral sensitivity analysis on the wave bands selected in the step (11), wherein one is a green water body in the center of the lake, and the other is a black green water body in a shallow water region, obtaining a wave band B3 with the highest spectral reflectivity of the water body and a second highest wave band B2, and setting the spectral information values corresponding to the Bi wave bands as rhoiAnd determining the initial combination of wave bands of the highlighted water body in the remote sensing image as k rho3+ρ2And k is the undetermined weight value.
Further, the step (2) includes:
(21) classifying the land features in Landsat7ETM + and Landsat 8OLI remote sensing images, carrying out statistical analysis on the influence degree of different land features on water body extraction, and finally determining five types of land features with the highest influence on water body extraction as high-reflection buildings, low-reflection buildings, vegetation, bare land and shadow;
(22) and (3) carrying out spectral sensitivity analysis on the typical feature selected in the step (21) facing to the common waveband of the screened Landsat7ETM + and Landsat 8OLI remote sensing images according to the difference of spectral information to obtain the spectral reflectivities of the typical feature on wavebands B5, B6, B7 and B4 which are sequentially decreased, and setting the spectral information values corresponding to the Bi waveband as rhoiIn order to better distinguish the difference between the water body and the sensitive wave band combination of five typical ground objects as rho4+ρ5+ρ6+ρ7。
Further, the method for determining the band weight based on the idea of the multi-objective optimization algorithm in the step (3) includes:
(31) carrying out recombination optimization on the sensitive wave bands of the water body obtained in the step (1) and the step (2) and the sensitive wave bands obtained by five typical objects, and utilizing k rho3+ρ2And rho4+ρ5+ρ6+ρ7Constructing a final water body extraction model according to the size relationship; the formalized description of the established optimization problem is as follows:
max∑x∈Xkρ3x+ρ2x-m(ρ4x+ρ5x+ρ6x+ρ7x)
so that
Figure BDA0002781789210000031
min∑x∈Ykρ3x+ρ2x-m(ρ4x+ρ5x+ρ6x+ρ7x)
So that
Figure BDA0002781789210000032
Wherein X represents a water body region, Y represents five typical feature regions, rho ixRepresenting the spectral information value of the pixel x corresponding to the Bi wave band, wherein k and m are undetermined weight values;
(32) solving an integer solution to the optimization problems (1) and (2) using a strategy that defines a variable range re-solution, comprising: when k is more than or equal to 1 and less than 3 and m is more than or equal to 1 and less than 3, calculating the wave band information of each pixel based on random sampling, and calculating the wave band combination value k rho corresponding to two types of water bodies and five types of typical ground objects3+ρ2-m(ρ4+ρ5+ρ6+ρ7) The result is that all the wave band combination values calculated by all the water bodies are all larger than 0, and the wave band combination values obtained by the five types of typical objects are all smaller than 0; when k is equal to or greater than 3, if m is equal to 1, k ρ is calculated based on random sampling3+ρ2-m(ρ4+ρ5+ρ6+ρ7) The result is that the wave band combination values corresponding to the water body and the typical ground object are all larger than 0, so that the water body and the non-water body are not favorably distinguished; determining the value range of k to be more than or equal to 1 and less than or equal to k and less than 3, and analyzing to obtain that when the value of k is more than or equal to 1 and less than or equal to 3, the difference of the wave band combination values of the water body and the bare soil is gradually reduced along with the increase of the value of k; and in the range that k is more than or equal to 1 and less than 3, analyzing the distribution condition of the band combination values obtained by different weight combination calculation, thereby determining the optimal solution meeting the two optimization problems.
Further, the band characteristic-based Landsat remote sensing image water body extraction model constructed in the step (3) is as follows:
NMBWI＝2ρ3+ρ2-(ρ4+ρ5+ρ6+ρ7) (3)
has the advantages that: according to the method for extracting the water body of the Landsat remote sensing image based on the wave band characteristics, the construction of a water body extraction model is carried out on the sensitivities of six common wave bands in Landsat7ETM + and Landsat 8OLI remote sensing images through the spectral information of water bodies and other typical ground objects according to the spectral statistical characteristics of actual remote sensing image data, so that the effective distinguishing of different ground objects is achieved, and the water body extraction precision is improved. Compared with the traditional water body extraction model, the method has stronger pertinence and robustness, can be simultaneously suitable for Landsat7ETM + and Landsat 8OLI remote sensing images, can more effectively distinguish water bodies and non-water body ground objects, can extract the water body by a stable threshold value of 0, and avoids the trouble that the traditional water body extraction model is difficult to determine the optimal threshold value.
Drawings
In order that the present disclosure may be more readily understood, reference is made to the following detailed description taken in conjunction with the accompanying drawings,
wherein:
fig. 1 is a flowchart of a water body extraction method according to an embodiment of the present invention.
FIG. 2 is a comparison graph of spectral reflectance analysis of various exemplary features identified in an example of the present invention.
FIG. 3 is a distribution diagram of calculated values corresponding to different combinations according to an embodiment of the present invention.
FIG. 4 is a comparison graph of different algorithm extraction results of the Taihu lake region.
FIG. 5 is a comparison graph of the extraction results of the nested lake region.
FIG. 6 is a comparison graph of the Poyang lake area extraction results.
Detailed Description
The present invention is further illustrated by the following figures and specific examples, which are to be understood as illustrative only and not as limiting the scope of the invention, which is to be given the full breadth of the appended claims and any and all equivalent modifications thereof which may occur to those skilled in the art upon reading the present specification.
As shown in fig. 1, the method for extracting the Landsat remote sensing image water body based on the band characteristics disclosed in the embodiment of the present invention mainly includes the following steps:
selecting six common wave bands in Landsat7ETM + and Landsat 8OLI remote sensing images, carrying out spectrum sensitivity analysis on the common wave bands of two types of water bodies in the center of a lake and a shallow water area one by one, determining the selection of the sensitive wave bands of each type of object by utilizing the spectrum information difference of the same object on different wave bands, and giving an initial weight (the initial weight can be set to be 1) to the selected wave bands. The method specifically comprises the following steps:
(11) screening six common wave bands of Landsat7ETM + and Landsat 8OLI remote sensing images, wherein the six common wave bands comprise Blue (B2), Green (B3), Red (B4), NIR (B5), SWIR-1(B6) and SWIR-2(B7) wave bands, and determining specific water body classes and typical object classes of spectral sensitivity analysis;
(12) and (3) carrying out spectral sensitivity analysis on the wave bands selected in the step (11), wherein one is a green water body in the center of the lake, and the other is a black green water body in a shallow water region, obtaining a wave band B3 with the highest spectral reflectivity of the water body and a second highest wave band B2, and setting the spectral information values corresponding to the Bi wave bands as rhoiAnd determining the initial combination of wave bands of the highlighted water body in the remote sensing image as k rho3+ρ2And k is the undetermined weight value.
And secondly, performing spectral sensitivity analysis on five typical land features, namely high-reflection buildings, low-reflection buildings, vegetation, bare land and shadow, except the water body aiming at the common wave band of the Landsat7ETM + and Landsat 8OLI remote sensing images, and determining the common sensitive wave band of the five land features and the initial weight corresponding to the wave band by combining the sensitive wave band of the water body. The method specifically comprises the following steps:
(21) classifying the land features in Landsat7ETM + and Landsat 8OLI remote sensing images, carrying out statistical analysis on the influence degree of different land features on water body extraction, and determining five types of land features with the highest influence on water body extraction as high-reflection buildings, low-reflection buildings, vegetation, bare land and shadow;
(22) and (3) carrying out spectral sensitivity analysis on the typical feature selected in the step (21) according to the difference of spectral information for the common waveband of the screened Landsat7ETM + and Landsat 8OLI remote sensing images to obtain the typical feature at wavebands B5, B6 and B7, the spectral reflectivity on the B4 is sequentially decreased, and the spectral information values corresponding to the Bi wave band are respectively rhoiIn order to better distinguish the difference between the water body and the sensitive wave band combination of five typical ground objects as rho4+ρ5+ρ6+ρ7(ii) a The results of the spectral reflectance analysis of the water body and five typical features are shown in fig. 2.
And step three, based on the idea of a multi-objective optimization algorithm, on the principle of better distinguishing two types of water bodies and five types of typical ground objects, combining the sensitive wave bands of the water bodies and the common sensitive wave bands of the five types of typical ground objects, determining the weight of the wave bands, and constructing a water body extraction model, which specifically comprises the following steps:
(31) the sensitive wave bands of the water body obtained in the first two steps and the sensitive wave bands obtained by the five typical objects are recombined and optimized, and k rho is utilized3+ρ2And rho4+ρ5+ρ6+ρ7Constructing a final water body extraction model according to the size relationship; the formalized description of the established optimization problem is as follows:
max∑x∈Xkρ3x+ρ2x-m(ρ4x+ρ5x+ρ6x+ρ7x)
so that
Figure BDA0002781789210000051
min∑x∈Ykρ3x+ρ2x-m(ρ4x+ρ5x+ρ6x+ρ7x)
So that
Figure BDA0002781789210000052
Wherein X represents a water body region, Y represents five typical feature regions, rho ixRepresenting the spectral information value of the pixel x corresponding to the Bi wave band, wherein k and m are undetermined weight values;
(32) solving an integer solution to the optimization problems (1) and (2) using a strategy that defines a variable range re-solution, comprising:
when k is more than or equal to 1 and less than 3 and m is more than or equal to 1 and less than 3, based onRandomly sampling and calculating the wave band information of each pixel, and calculating the wave band combination value k rho corresponding to two types of water bodies and five types of typical ground objects3+ρ2-m(ρ4+ρ5+ρ6+ρ7) The result is that all the wave band combination values calculated by all the water bodies are all larger than 0, and the wave band combination values obtained by the five types of typical objects are all smaller than 0;
when k is equal to or greater than 3, if m is equal to 1, k ρ is calculated based on random sampling3+ρ2-m(ρ4+ρ5+ρ6+ρ7) The result is that the wave band combination values corresponding to the water body and the typical ground object are all larger than 0, so that the water body and the non-water body are not favorably distinguished; determining the value range of k to be more than or equal to 1 and less than or equal to k and less than 3, and analyzing to obtain that when the value of k is more than or equal to 1 and less than or equal to 3, the difference of the wave band combination values of the water body and the bare soil is gradually reduced along with the increase of the value of k;
and in the range that k is more than or equal to 1 and less than 3, analyzing the distribution condition of the band combination values obtained by different weight combination calculation, thereby determining the optimal solution meeting the two optimization problems.
In this embodiment, the calculated values of the terrestrial objects obtained by different weights under the above k constraint condition are shown in fig. 3. In this embodiment, k is determined to be 2, m is 1, and the finally constructed Landsat remote sensing image Water body extraction model based on the Band characteristics is (novel multiband Water body Index method, New Multi-Band Water Index):
NMBWI＝2ρ3+ρ2-(ρ4+ρ5+ρ6+ρ7) (3)
setting a threshold value based on the water body extraction model to realize binary segmentation and extraction of the water body, and specifically comprising the following steps of:
(41) carrying out spectrum value calculation again on the remote sensing image by using the constructed water body extraction model, and fusing multiband information of one pixel into one spectrum information through a band combination calculation formula, thereby obtaining a multiband fused remote sensing image based on the water body extraction model;
(42) the water body and the non-water body are distinguished by a threshold value of 0. The calculated value of the water body obtained according to the water body extraction model is larger than 0, and the calculated value of the non-water body is smaller than 0, so that the threshold value is set to be 0, binary segmentation of the water body is carried out on the basis of the fused spectral information of the pixels and the threshold value, the pixels with the spectral values larger than 0 are the water body, the pixels with the spectral values smaller than 0 are the non-water body, binary images are generated on the basis of the binary segmentation, and the final optimized water body extraction result is completed.
The NMBWI and other algorithm models are subjected to comparative experiments and result analysis, and the extraction results of the experiments are shown in FIG. 4.
(1) Subjective evaluation
As shown in FIG. 4, the four result graphs are compared and analyzed with the original image by visual interpretation, and generally, the four methods can basically extract the water body of the Taihu lake, so that the water and land separation is well realized.
The single-band threshold method is simple to operate, but a non-water body area can be extracted by mistake; the NDWI ratio model does not consider low-reflection ground objects during construction, and a shadow area can be extracted by mistake; the maximum likelihood method omits more tiny water body areas; the NMBWI method emphasizes considering the water body and low-reflection ground objects in a shallow water area during construction, and the extracted area is the closest to the actual water body area.
(2) Objective evaluation
200 water body sample points and 360 non-water body sample points are randomly selected from the Taihu lake image, the Google Earth high-resolution image water body is used as a reference picture, the extraction precision of each water body extraction algorithm is measured by using a precision evaluation index, and the result is shown in Table 1.
TABLE 1 comparison of water extraction results of different methods in Landsat 8 image Taihu lake region
Figure BDA0002781789210000071
Analysis of different area experiment comparison results
(1) Subjective evaluation
And selecting water bodies in different areas to further test the robustness of the method. FIG. 5 is a plot of the extraction results of the water bodies in the nested lake, and FIG. 6 is a plot of the extraction results of the water bodies in the Poyang lake, and it can be known through visual interpretation of the water body extraction results in different areas that the method of the present invention can better eliminate shadows compared with NDWI, and has a good extraction effect for extracting large-area lakes with large water body areas in shallow water areas along the shore, and a certain robustness.
(2) Objective evaluation
Randomly selecting 80 water sample points and 120 non-water sample points from the nested lake image, randomly selecting 90 water sample points and 170 non-water sample points from the Poyang lake image, taking GoogleEarth high-resolution image water as a reference picture, and measuring the extraction precision of each water extraction algorithm by using a precision evaluation index, wherein the result is shown in a table 2.
TABLE 2 comparison of water extraction results of different methods in different regions of Landsat 8 image
Figure BDA0002781789210000072
As can be seen from the data in the table 2, the overall accuracy and Kappa coefficient of the NMBWI method in the water extraction of the two regions are higher than those of the NDWI method, the extraction result of the NMBWI is more consistent with the GoogleEarth image with higher resolution, and the robustness of the NMBWI method is verified.

Claims (5)

1. A method for extracting a Landsat remote sensing image water body based on wave band characteristics is characterized by comprising the following steps:
(1) selecting six common wave bands in Landsat7ETM + and Landsat 8OLI remote sensing images, carrying out spectrum sensitivity analysis on two types of water bodies of a lake center and a shallow water area one by one on the common wave bands, determining the selection of the sensitive wave bands of each type of object by utilizing the spectrum information difference of the same object on different wave bands and giving an initial weight to the selected wave bands;
(2) spectral sensitivity analysis of five typical types of ground objects, namely high-reflection buildings, low-reflection buildings, vegetation, bare land and shadow, except a water body is carried out on the common wave bands of Landsat7ETM + and Landsat 8OLI remote sensing images, and the common sensitive wave bands of the five types of ground objects and initial weights corresponding to the wave bands are determined by combining the sensitive wave bands of the water body;
(3) based on the idea of a multi-objective optimization algorithm, a principle of better distinguishing two types of water bodies and five types of typical ground objects is used, sensitive wave bands of the water bodies and common sensitive wave bands of the five types of typical ground objects are combined, weights of the wave bands are determined, and a water body extraction model is constructed;
(4) calculating new spectral information of each pixel point according to the water body extraction model, thereby obtaining a multiband fused remote sensing image based on the water body extraction model;
(5) and according to the difference of the calculated values of the water body extraction model, taking 0 as a threshold value, distinguishing the water body and the non-water body by threshold value segmentation, taking the water body with the spectral information larger than 0 and the non-water body with the spectral information smaller than 0, and generating a binary image according to a binary segmentation result to realize water body extraction.
2. The method for extracting the Landsat remote sensing image water body based on the wave band characteristics as claimed in claim 1, wherein the step (1) comprises:
(11) selecting six common wave bands B2-B7 of Landsat7ETM + and Landsat 8OLI remote sensing images, respectively corresponding to Blue, Green, Red, NIR, SWIR-1 and SWIR-2 wave bands, and determining specific water body types and typical object types for spectral sensitivity analysis;
(12) and (3) carrying out spectral sensitivity analysis on the wave bands selected in the step (11), wherein one is a green water body in the center of the lake, and the other is a black green water body in a shallow water region, obtaining a wave band B3 with the highest spectral reflectivity of the water body and a second highest wave band B2, and setting the spectral information values corresponding to the Bi wave bands as rhoiAnd determining the initial combination of wave bands of the highlighted water body in the remote sensing image as k rho3+ρ2And k is the undetermined weight value.
3. The method for extracting the Landsat remote sensing image water body based on the wave band characteristics as claimed in claim 1, wherein the step (2) comprises:
(21) classifying the land features in Landsat7ETM + and Landsat 8OLI remote sensing images, carrying out statistical analysis on the influence degree of different land features on water body extraction, and finally determining five types of land features with the highest influence on water body extraction as high-reflection buildings, low-reflection buildings, vegetation, bare land and shadow;
(22) for the typical feature selected in step (21), faceSpectrum sensitivity analysis is carried out on the common wave bands of the screened Landsat7ETM + and Landsat 8OLI remote sensing images according to the difference of the spectrum information to obtain the spectrum reflectivity of the typical objects on the wave bands B5, B6, B7 and B4 which are sequentially decreased, and the spectrum information values corresponding to the Bi wave band are respectively rhoiIn order to better distinguish the difference between the water body and the sensitive wave band combination of five typical ground objects as rho4+ρ5+ρ6+ρ7。
4. The method for extracting the Landsat remote sensing image water body based on the band characteristics as claimed in claim 1, wherein the method for determining the band weight based on the multi-objective optimization algorithm in step (3) comprises:
(31) carrying out recombination optimization on the sensitive wave bands of the water body obtained in the step (1) and the step (2) and the sensitive wave bands obtained by five typical objects, and utilizing k rho3+ρ2And rho4+ρ5+ρ6+ρ7Constructing a final water body extraction model according to the size relationship; the formalized description of the established optimization problem is as follows:
max∑x∈Xkρ3x+ρ2x-m(ρ4x+ρ5x+ρ6x+ρ7x)
so that
Figure FDA0002781789200000021
min∑x∈Ykρ3x+ρ2x-m(ρ4x+ρ5x+ρ6x+ρ7x)
So that
Figure FDA0002781789200000022
Wherein X represents a water body region, Y represents five typical feature regions, rho ixRepresenting the spectral information value of the pixel x corresponding to the Bi wave band, wherein k and m are undetermined weight values;
(32) solving optimization problems (1) and (2) using a strategy that defines a variable range to solveComprises: when k is more than or equal to 1 and less than 3 and m is more than or equal to 1 and less than 3, calculating the wave band information of each pixel based on random sampling, and calculating the wave band combination value k rho corresponding to two types of water bodies and five types of typical ground objects3+ρ2-m(ρ4+ρ5+ρ6+ρ7) The result is that all the wave band combination values calculated by all the water bodies are all larger than 0, and the wave band combination values obtained by the five types of typical objects are all smaller than 0;
when k is equal to or greater than 3, if m is equal to 1, k ρ is calculated based on random sampling3+ρ2-m(ρ4+ρ5+ρ6+ρ7) The result is that the wave band combination values corresponding to the water body and the typical ground object are all larger than 0, so that the water body and the non-water body are not favorably distinguished; determining the value range of k to be more than or equal to 1 and less than or equal to k and less than 3, and analyzing to obtain that when the value of k is more than or equal to 1 and less than or equal to 3, the difference of the wave band combination values of the water body and the bare soil is gradually reduced along with the increase of the value of k; and in the range that k is more than or equal to 1 and less than 3, analyzing the distribution condition of the band combination values obtained by different weight combination calculation, thereby determining the optimal solution meeting the two optimization problems.
5. The method for extracting the water body from the Landsat remote sensing image based on the band characteristics as claimed in claim 4, wherein the Landsat remote sensing image water body extraction model based on the band characteristics constructed in the step (3) is as follows:
NMBWI＝2ρ3+ρ2-(ρ4+ρ5+ρ6+ρ7) (3)。
CN202011284176.6A 2020-11-17 2020-11-17 Landsat remote sensing image water body extraction method based on wave band characteristics Active CN112270301B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011284176.6A CN112270301B (en) 2020-11-17 2020-11-17 Landsat remote sensing image water body extraction method based on wave band characteristics

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011284176.6A CN112270301B (en) 2020-11-17 2020-11-17 Landsat remote sensing image water body extraction method based on wave band characteristics

Publications (2)

Publication Number Publication Date
CN112270301A true CN112270301A (en) 2021-01-26
CN112270301B CN112270301B (en) 2022-07-29

Family

ID=74340566

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011284176.6A Active CN112270301B (en) 2020-11-17 2020-11-17 Landsat remote sensing image water body extraction method based on wave band characteristics

Country Status (1)

Country Link
CN (1) CN112270301B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112800823A (en) * 2021-04-19 2021-05-14 上海普适导航科技股份有限公司 Rapid water body extraction method

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104751179A (en) * 2015-04-01 2015-07-01 河海大学 Multi-target high spectral remote sensing image wave band selection method based on game theory
CN110033499A (en) * 2019-04-21 2019-07-19 南京林业大学 A kind of regional water body drafting method based on Landsat OLI remote sensing image

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104751179A (en) * 2015-04-01 2015-07-01 河海大学 Multi-target high spectral remote sensing image wave band selection method based on game theory
CN110033499A (en) * 2019-04-21 2019-07-19 南京林业大学 A kind of regional water body drafting method based on Landsat OLI remote sensing image

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112800823A (en) * 2021-04-19 2021-05-14 上海普适导航科技股份有限公司 Rapid water body extraction method
CN112800823B (en) * 2021-04-19 2024-04-02 上海普适导航科技股份有限公司 Water body rapid extraction method

Also Published As

Publication number Publication date
CN112270301B (en) 2022-07-29

Similar Documents

Publication Publication Date Title
CN109993237B (en) Water body rapid extraction method and system based on high-resolution satellite optical remote sensing data
Ouma et al. A water index for rapid mapping of shoreline changes of five East African Rift Valley lakes: an empirical analysis using Landsat TM and ETM+ data
CN107330875B (en) Water body surrounding environment change detection method based on forward and reverse heterogeneity of remote sensing image
CN111274865A (en) Remote sensing image cloud detection method and device based on full convolution neural network
CN109726705B (en) Mangrove forest information extraction method and device and electronic equipment
CN111985543A (en) Construction method, classification method and system of hyperspectral image classification model
CN108629287A (en) A kind of remote sensing image terrain classification method
CN107230197B (en) Tropical cyclone objective strength determination method based on satellite cloud image and RVM
CN111192260B (en) Melon quality detection method based on hyperspectral image depth feature fusion
Almeer Vegetation extraction from free google earth images of deserts using a robust BPNN approach in HSV Space
CN109598202A (en) A kind of object-based satellite image multi objective built-up areas extraction method
CN114494821A (en) Remote sensing image cloud detection method based on feature multi-scale perception and self-adaptive aggregation
CN110570462A (en) flood inundation range automatic extraction method based on polarized radar remote sensing image
CN114778483A (en) Method for correcting terrain shadow of remote sensing image near-infrared wave band for monitoring mountainous region
Namikawa et al. Water body extraction from Rapideye images: An automated methodology based on hue component of color transformation from RGB to HSV model
CN112270301B (en) Landsat remote sensing image water body extraction method based on wave band characteristics
CN113538559A (en) Extraction method of offshore aquaculture raft extraction index based on hyperspectral remote sensing image
CN114021656A (en) Water body extraction method based on GEE cloud platform and optical and radar data fusion
Liang et al. Automatic remote sensing detection of floating macroalgae in the yellow and east china seas using extreme learning machine
He et al. Object-based distinction between building shadow and water in high-resolution imagery using fuzzy-rule classification and artificial bee colony optimization
Lu et al. Application of red edge band in remote sensing extraction of surface water body: A case study based on GF-6 WFV data in arid area
CN108198178B (en) Method and device for determining atmospheric range radiation value
CN116994071A (en) Multispectral laser radar point cloud classification method based on self-adaptive spectrum residual error
CN111882573A (en) Cultivated land plot extraction method and system based on high-resolution image data
CN115953685A (en) Multi-layer multi-scale division agricultural greenhouse type information extraction method and system

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