CN117611888A - Water body classification method and device based on shape and water body submerged frequency characteristics - Google Patents
Water body classification method and device based on shape and water body submerged frequency characteristics Download PDFInfo
- Publication number
- CN117611888A CN117611888A CN202311567441.5A CN202311567441A CN117611888A CN 117611888 A CN117611888 A CN 117611888A CN 202311567441 A CN202311567441 A CN 202311567441A CN 117611888 A CN117611888 A CN 117611888A
- Authority
- CN
- China
- Prior art keywords
- water body
- water
- image data
- classification
- shape
- 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.)
- Pending
Links
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 287
- 238000000034 method Methods 0.000 title claims abstract description 71
- 239000002352 surface water Substances 0.000 claims abstract description 62
- 230000037237 body shape Effects 0.000 claims abstract description 33
- 238000007637 random forest analysis Methods 0.000 claims abstract description 25
- 238000003709 image segmentation Methods 0.000 claims abstract description 5
- 230000001932 seasonal effect Effects 0.000 claims description 46
- 238000004364 calculation method Methods 0.000 claims description 29
- 230000011218 segmentation Effects 0.000 claims description 9
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000011160 research Methods 0.000 abstract description 14
- 238000012544 monitoring process Methods 0.000 abstract description 6
- 238000011156 evaluation Methods 0.000 abstract description 4
- 230000002349 favourable effect Effects 0.000 abstract description 3
- 238000002310 reflectometry Methods 0.000 description 14
- 230000007613 environmental effect Effects 0.000 description 12
- 238000012937 correction Methods 0.000 description 10
- 238000000605 extraction Methods 0.000 description 9
- 241000209094 Oryza Species 0.000 description 6
- 235000007164 Oryza sativa Nutrition 0.000 description 6
- 235000009566 rice Nutrition 0.000 description 6
- 230000008859 change Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 238000006424 Flood reaction Methods 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 4
- 230000003595 spectral effect Effects 0.000 description 4
- 230000015572 biosynthetic process Effects 0.000 description 3
- 238000012952 Resampling Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000005520 cutting process Methods 0.000 description 2
- 230000001788 irregular Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000005855 radiation Effects 0.000 description 2
- 239000002689 soil Substances 0.000 description 2
- 238000001228 spectrum Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000000746 body region Anatomy 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 239000008239 natural water Substances 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 238000003809 water extraction Methods 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/01—Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Image Processing (AREA)
Abstract
The invention relates to a water body classification method based on shape and water body inundation frequency characteristics, belonging to the technical field of water resource research, comprising the following steps: and acquiring remote sensing image data, performing image segmentation on the remote sensing image data to obtain water body image data, extracting surface water body shape characteristic information and water body inundation frequency information based on the water body image data, and classifying surface water bodies by using a random forest classifier based on the water body shape characteristic information and the water body inundation frequency information. The scheme carries out fine classification on the surface water body based on the shape characteristics and the water body inundation frequency characteristics, realizes more accurate water body classification, and provides favorable technical support for water resource management and accurate monitoring and evaluation of ecological environment.
Description
Technical Field
The invention relates to the technical field of water resource research, in particular to a water body classification method and device based on shape and water body inundation frequency characteristics.
Background
The water classification research has important significance for water environmental protection and industrial production.
Many remote sensing data sets related to surface water bodies at present lack finer classification, and most of the remote sensing data sets only take the surface water bodies as single surface coverage types, so that a water body classification system is not comprehensive enough and cannot provide complete surface water body type information.
In addition, many studies use methods of manual identification, manual editing for surface water body data set production, which is inefficient, time consuming and costly.
Accordingly, there is a need for a method of water classification based on shape characteristics and water flooding frequency characteristics that overcomes the above-described problems.
Disclosure of Invention
One of the purposes of the present application is to provide a method and a device for classifying water based on shape and water flooding frequency characteristics, so as to solve the problems in the prior art.
According to an embodiment of an aspect of the present invention, there is provided a water classification method based on a shape t and a water flooding frequency characteristic, including:
acquiring remote sensing image data;
carrying out water body index calculation on the remote sensing image data, and dividing a water body index calculation result to obtain water body image data;
extracting surface water body shape characteristic information and water body inundation frequency information based on the water body image data;
and classifying the surface water body by using a random forest classifier based on the water body shape characteristic information and the water body inundation frequency information.
Preferably, the step of dividing the water body index calculation result to obtain water body image data includes:
threshold segmentation is carried out on the water index calculation result by using an Ojin method, so as to obtain a water-background two-classification chart;
carrying out image gray threshold segmentation on the two classification images, and separating a water body from a background;
and dividing the separated water body by using a multi-scale dividing method to obtain water body image data.
Preferably, the step of classifying the surface water body by using the random forest classifier based on the water body shape characteristic information and the water body inundation frequency information comprises the following steps:
performing first-layer classification on the surface water body by using a random forest classifier based on the water body shape characteristic information;
and carrying out second-layer classification on the surface water body by utilizing a random forest classifier based on the water body inundation frequency information.
Preferably, the first layer classification is used for distinguishing that the water body category belongs to one of first preset categories, and the first preset categories include: lakes, rivers, reservoirs, ponds, paddy fields and ditches.
Preferably, the second layer classification is used for distinguishing that the water body category belongs to one of second preset categories, and the second preset categories include: permanent water, seasonal water, and maximum water.
Preferably, before extracting the surface water body shape feature information based on the water body image data, the method further includes:
shape features are selected from a number of surface body geometric features by feature importance calculations.
Preferably, the selected shape features include:
compactness, area, length, boundary index, and shape index;
and extracting surface water body shape characteristic information based on the water body image data to obtain the selected shape characteristic information.
Preferably, the water flooding frequency information is obtained by superposing water graphs at different times, and the water flooding frequency WIF of each pixel in the whole time sequence is calculated by the following calculation method:
n represents the number of pixel values for all good observations,W ij is binary changeQuantity, representing imagejMiddle pixeli1 for flooding, 0 for non-flooding,O ij representing an imagejMiddle pixeli1 indicates observation, and 0 indicates no observation.
Preferably, before the remote sensing image data is subjected to image segmentation, the method further comprises:
preprocessing and correcting the remote sensing image data.
According to one embodiment of another aspect of the present invention, there is provided a water classification device based on shape and water flooding frequency characteristics, comprising:
the first unit is used for acquiring remote sensing image data;
the second unit is used for carrying out water body index calculation on the remote sensing image data and dividing a water body index calculation result to obtain water body image data;
the third unit is used for extracting surface water body shape characteristic information and water body inundation frequency information based on the water body image data;
and the fourth unit is used for classifying the surface water body by utilizing a random forest classifier based on the water body shape characteristic information and the water body inundation frequency information.
The invention has the beneficial effects that:
according to the embodiment of the invention, the remote sensing image data are acquired, the remote sensing image data are subjected to image segmentation to obtain the water body image data, and the surface water body shape characteristic information and the water body inundation frequency information are extracted based on the water body image data, so that the surface water body is classified by using a random forest classifier based on the water body shape characteristic information and the water body inundation frequency information. The scheme carries out fine classification on the surface water body based on the shape characteristics and the water body inundation frequency characteristics, realizes more accurate water body classification, and provides favorable technical support for water resource management and accurate monitoring and evaluation of ecological environment.
Those of ordinary skill in the art will realize that while the following detailed description proceeds with reference to the illustrative embodiments, the accompanying drawings, the present application is not limited to only these embodiments. Rather, the scope of the present application is broad and is intended to be limited only by the claims appended hereto.
Drawings
Other features, objects and advantages of the present application will become more apparent upon reading of the detailed description of non-limiting embodiments, made with reference to the following drawings, in which:
FIG. 1 is a flow chart of a method for classifying water based on shape and water flooding frequency characteristics according to an embodiment of the present invention;
FIG. 2 is a diagram of surface water type classification according to an embodiment of the invention;
FIG. 3 is a comparative view of lake classification results according to an embodiment of the present invention;
FIG. 4 is a comparison of river classification results according to an embodiment of the present invention;
FIG. 5 is a comparison of reservoir classification results according to an embodiment of the invention;
FIG. 6 is a graph showing a comparison of paddy field and pond classification results according to an embodiment of the present invention;
fig. 7 is a schematic diagram of a framework structure of a water body classification device according to an embodiment of the invention.
Those of ordinary skill in the art will realize that while the following detailed description proceeds with reference to the illustrative embodiments, the accompanying drawings, the present application is not limited to only these embodiments. Rather, the scope of the present application is broad and is intended to be limited only by the claims appended hereto.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The terms first, second and the like in the description and in the claims, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged, as appropriate, such that embodiments of the present invention may be implemented in sequences other than those illustrated or described herein, and that the objects identified by "first," "second," etc. are generally of a type, and are not limited to the number of objects, such as the first object may be one or more. Furthermore, the term "and/or" as used in the specification and claims to describe an association of associated objects means that there may be three relationships, e.g., a and/or B, may mean: a exists alone, A and B exist together, and B exists alone. The character "/" generally indicates that the context-dependent object is an "or" relationship. The term "plurality" in embodiments of the present invention means two or more, and other adjectives are similar.
Before discussing exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although a flowchart depicts operations as a sequential process, many of the operations can be performed in parallel, concurrently, or at the same time. Furthermore, the order of the operations may be rearranged. The process may be terminated when its operations are completed, but may have additional steps not included in the figures. The processes may correspond to methods, functions, procedures, subroutines, and the like.
The methods discussed below (some of which are illustrated by flowcharts) may be implemented by hardware, software, firmware, middleware, microcode, hardware description languages, or any combination thereof. When implemented in software, firmware, middleware or microcode, the program code or code segments to perform the necessary tasks may be stored in a machine or computer readable medium such as a storage medium. The processor(s) may perform the necessary tasks.
The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and/or "comprising," when used herein, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
It should also be noted that, in some alternative implementations, the functions/acts noted may occur out of the order noted in the figures. For example, two figures shown in succession may in fact be executed substantially concurrently or the figures may sometimes be executed in the reverse order, depending upon the functionality/acts involved.
In the research process, the inventor finds that many remote sensing data sets related to the surface water body lack finer classification, and the surface water body elements, global surface water body thematic data sets and the like in the whole-ball land coverage remote sensing data products are mostly only used as single surface coverage types, so that the water body classification system is not comprehensive enough, complete surface water body type information cannot be provided, for example, the water body types such as lakes, rivers and reservoirs cannot be independently identified, and seasonal change information of the surface water body cannot be reflected. In addition, many studies use methods of manual identification, manual editing for surface water body data set production, which is inefficient, time consuming and costly.
Therefore, the limitation that the existing method is difficult to realize the fine classification of the surface water body in the complex environment is broken through, and the method for establishing the efficient and fine surface water body type identification is to meet the key requirements of the fine management of water resources and the accurate monitoring and evaluation of ecological environment.
Based on the characteristics, the application provides a water body classification method based on the shape and the water body submerged frequency characteristics, so as to solve the problems. The technical solutions of the present application are described in further detail below with reference to the accompanying drawings.
As shown in fig. 1, a flow chart of a water classification method based on shape and water flooding frequency features according to an embodiment of the present application is provided, and the method includes the following operations:
s10, acquiring remote sensing image data;
the method for acquiring remote sensing image data can be realized based on the existing optical satellite, for example, the sentry No. 2 (Sentinel-2) is a sentry series high-resolution optical satellite, and consists of two satellites of 2A and 2B, the revisiting period is 5 days, the revisiting period in a high-latitude area only needs 3 days, the revisiting requirement and the coverage requirement can be met, and the method is widely applied to global higher-resolution earth surface parameter inversion, repeated observation and disaster monitoring research. A multispectral imager (MSI) carried by a Sentinel-2 satellite comprises 13 spectral bands such as visible light, near infrared, short wave infrared and the like, the breadth of an image is 290km, and the spatial resolution is 10m, 20m and 60m. The remote sensing data used in the application is a Sentinel-2 Level 2A product from 1 st 2021 to 31 nd 2021, and the Level-2A image is a standard product of surface reflectivity and is subjected to orthocorrection of the bottom layer reflectivity (BOA) of the atmosphere.
In order to ensure the accuracy of data processing, the embodiment of the application performs preprocessing and correction on acquired remote sensing image data, wherein the preprocessing comprises: wave band combination, cloud cover removal, mosaic, cutting and the like; the correction includes: atmospheric correction, radiation correction, geometric correction and/or resampling, and the like.
S11, carrying out water body index calculation on the remote sensing image data, and dividing a water body index calculation result to obtain water body image data;
the method for calculating the water body index of the remote sensing image data can comprise the following steps:
1) Extracting environmental background characteristic information in the remote sensing image data;
the environmental background characteristic information includes, but is not limited to: normalized vegetation Index (Normalized Difference Vegetation Index, NDVI), corrected normalized water Index (Modified Normalized Difference Water Index, MNDWI), normalized building Index (NormalizedDifference Built-up Index, NDBI), and corrected bare soil Index (Modified Bare Soil Index, MBSI).
The method for extracting the environmental background characteristic information in the remote sensing image data comprises the following steps: firstly, calculating environmental background characteristics of the remote sensing image data; then, equally dividing the remote sensing images into 50 categories based on the value of each environmental background feature; and calculating the number of pixels in each category, wherein each environmental background feature generates a 50×1 environmental background feature information vector, and finally obtaining the statistical information superposition of four environmental background features to obtain a 200×1 environmental background feature information vector.
2) Performing similarity calculation on the environmental background characteristic information based on background environment priori knowledge to determine the environment type of the water body region to be analyzed;
the main purpose of the step is to determine the environment type through similarity calculation so as to construct a corresponding water body index based on the environment type, namely, the similarity between the image block of the water body area to be analyzed and the priori knowledge of the background environment is calculated through a background similarity calculation method, and then the environment type of the image block is judged to be an urban or non-urban area. The embodiment of the application can calculate the similarity between the background characteristics and the priori knowledge by using a cosine similarity algorithm so as to judge the type of the background environment. And selecting Sentinel-2 with a resolution of 10 meters as a data source, and dividing the whole image into non-overlapping sub-images according to a scale of 200m multiplied by 200 m. Then, an environmental background similarity function is constructed, and the similarity between each input regular image block and the priori knowledge is calculated.
3) And constructing a water body index corresponding to the determined environment type.
According to the embodiment of the application, through typical feature spectral feature analysis, the water body indexes applicable to different environments (cities/non-cities) are respectively constructed, and then the surface water body extraction under different environments is carried out by combining a background feature similarity algorithm. According to the spectral characteristic reflectivity of eight typical features under different environments, the method for calculating the constructed water body index applicable to cities and non-cities is as follows
（1）
（2）
（3）
（4）
Indicating red band reflectivity +.>Indicating the reflectivity of the blue band, ">Indicating the reflectivity of the green band, ">Indicating the reflectivity of the NIR (near infrared) band, +.>Indicating SWIR1 (short wave infrared 1) band reflectivity,representing the SWIR2 (short wave infrared 2) band reflectivity, < >>Representing Vegetation Red Edge (vegetation spectrum red edge) band reflectivity, < >>Indicating the reflectivity of the NIR narrow band,/for the NIR narrow spectrum>Values representing the maximum reflectance band in the bands Blue, green, red, NIR, SWIR1, SWIR2,/for the wavelength bands Blue, green, red, NIR, SWIR1, SWIR2>Representing the minimum reflectivity wave in the above wave bandValues of the segments;WEBI nc representing the index of the background water body of the non-city,WEBI c representing an urban background water index, and calculating the water index by using the formula (4) if the background similarity is greater than 0.5 and the environment type of the water area to be analyzed is the urban background; when the background similarity is smaller than 0.5, calculating the water index by using the formula (3) when the environment type of the water area to be analyzed is a non-urban background.
After obtaining the water body index, the embodiment of the application segments the index calculation result to obtain the water body image data, and the specific method can comprise the following steps:
firstly, threshold segmentation is carried out on a water index calculation result by using an Ojin method, and a (water-background) two-classification chart is obtained;
then, carrying out image gray threshold segmentation on the result, and separating the water body from the background, wherein a large number of adjacent water bodies of different types are connected together;
and finally, searching an optimal segmentation scale, and segmenting the result by using a multi-scale segmentation method to finally obtain an independent single type segmentation object, namely obtaining the water body image data.
S12, extracting surface water body shape characteristic information and water body inundation frequency information based on the water body image data;
the surface water body classification system is constructed based on the water body formation mode and the water body inundation frequency information, as shown in fig. 2, the first stage classification (first classification in fig. 2) can divide the surface water body into a natural water body, a semi-artificial water body and an artificial water body according to the surface water body formation mode; the second stage classification (classification two) is to divide the surface water body into the following according to the water body shape characteristics: lakes, rivers, reservoirs, paddy fields/ponds, and ditches; the third stage of water body (category three) is to divide the surface water body into the following according to the water body submerged frequency: seasonal wetlands, seasonal rivers, seasonal reservoirs, paddy fields, non-seasonal lakes, non-seasonal rivers, non-seasonal reservoirs, ponds, ditches, and the like.
Different types of surface waters are geometrically different, for example lakes are generally irregular polygons with larger areas, ponds and paddy fields are mostly regular rectangles with smaller areas, rivers are linear water bodies with long and thin curves, reservoirs are complex polygons with curved boundaries, and therefore the water shape characteristics can be selected to primarily distinguish the five categories of water bodies.
It will be appreciated that the body of water shape includes a number of features from which the body of water shape features used in the present invention need to be selected, and that the embodiments of the present application select shape features from a number of surface body of water geometry features by means of feature importance calculations. For example, 10 commonly used water body Shape features are firstly selected as classification initial features, 5 features are screened from the 10 selected features through feature importance calculation and are combined into a component feature space, and the component feature space is classified according to importance sequences, namely Compactness, area, length, boundary index and Shape index.
The method and the device perform judgment and classification on permanent water bodies (namely water body inundation frequency) by calculating the occurrence frequency of water body pixels in the images (namely water bodies which cannot change along with seasons and continuously exist all the year round, namely the minimum inundation range of the water bodies), and on quaternary water bodies (namely the water bodies with the water body inundation range which can change along with seasons) and maximum water bodies (namely the maximum inundation range of the water bodies). The water flooding frequency (WIF) refers to the ratio of the number of water floods that occur in a region to the total number of floods that occur in a time series. The water body submerged frequency information is obtained by superposing water body diagrams at different times, and the water body submerged frequency of each pixel in the whole time sequence is calculated by the following steps:
n represents the number of pixel values for all good observations,W ij is a binary variable representing an imagejMiddle pixeli1 for flooding, 0 for non-flooding,O ij representing an imagejMiddle pixeli1 indicates observation, and 0 indicates no observation.
S13, classifying the surface water body by using a random forest classifier based on the water body shape characteristic information and the water body inundation frequency information.
According to the water body type constructed by the method, different water bodies of different types have different shape characteristics and water body submerged frequency characteristics, so that the WCSF (Water Classification based on Shape features and Flooding frequency) method based on the shape characteristics and the water body submerged frequency characteristics is constructed for carrying out surface water body refinement type identification.
The shape characteristics of different types of water bodies are different, and firstly, the random forest classifier based on the shape characteristics can realize the identification of five types of water bodies in lakes, rivers, reservoirs, ponds/paddy fields and ditches. Secondly, the non-seasonal lake, the non-seasonal river and the non-seasonal water reservoir are permanent water bodies, the water bodies are covered almost all the year round, the seasonal wetland, the seasonal river and the seasonal water reservoir are seasonal change water bodies, the water bodies are covered only in the water-rich period, and the dead water period is the land, so that the identification of the type can be realized through a random forest classifier based on the water body submerged frequency characteristics.
Therefore, the step of classifying the surface water body by using the random forest classifier based on the water body shape characteristic information and the water body inundation frequency information in the embodiment of the application comprises the following two steps:
1) Performing first-layer classification on the surface water body by using a random forest classifier based on the water body shape characteristic information; the first layer classification is used for distinguishing that the water body category belongs to one of first preset categories, and the first preset categories comprise: lakes, rivers, reservoirs, ponds, paddy fields and ditches.
2) And carrying out second-layer classification on the surface water body by utilizing a random forest classifier based on the water body inundation frequency information. The second layer classification is used for distinguishing that the water body category belongs to one of second preset categories, and the second preset categories comprise: permanent water, seasonal water, and maximum water.
The method and the device realize the refined water type identification, firstly identify the second class based on the shape characteristics, then identify the third class based on the water flooding frequency characteristics, and determine the first class according to the type corresponding relation in fig. 2.
In order to better demonstrate the water classification effect of the present application, several sets of experimental data are described below.
And selecting 500 non-quaternary lakes, seasonal wetlands, non-seasonal rivers, non-quaternary reservoirs, ponds, paddy fields and ditches in the research area, wherein the total number of the nine water verification samples is 4500. Wherein the quaternary water body is selected from the period of 7-9 months of high water, and the non-quaternary water body is selected from the period of 11-12 months of non-rainy season. In order to better analyze the accuracy of the WCSF classification result of the water classification method of the present application, the inventors selected four remote sensing data products related to the surface water type to compare with the extraction result herein. Because the area of the artificial water body ditch in the research area is smaller, and few researches are conducted on the ditch for extraction research, only lakes, rivers, reservoirs, paddy fields and ponds are selected for classification detail comparison analysis with the existing data set.
The overall accuracy of water body type identification by using the method disclosed by the application is 89.87%, and the kappa coefficient is 0.84. The classification accuracy of non-seasonal bodies of water is generally higher than that of seasonal bodies of water. The classification accuracy of non-seasonal lakes, non-seasonal rivers, non-seasonal reservoirs, paddy fields, ponds and ditches is over 90 percent, wherein the classification accuracy of the non-seasonal lakes is highest, the Precision (PA) of a producer is 92.46 percent, and the precision (UA) of a user is 91.07 percent; secondly, the river is a non-seasonal river, the precision of a producer is 92.02%, and the precision of a user is 91.58%; next, a non-quaternary water reservoir was produced with a producer precision of 91.58% and a user precision of 91.25%. The precision of producers and users of artificial water bodies (paddy fields, ponds and ditches) is higher than 90 percent. The pond producer precision is 91.37%, the user precision is 91.54%, the ditch producer precision is 90.87%, the user precision is 91.67%, the paddy field producer precision is 90.05%, and the user precision is 90.38%. The lowest classification precision is the seasonal wetland, the precision of a producer is 86.47%, and the precision of a user is 85.92%; secondly, the precision of a producer is 88.71% and the precision of a user is 89.16% in seasonal river; finally, the precision of the producer is 89.87% and the precision of the user is 88.20% in the quaternary water reservoir. Nine types of water classification accuracy are shown in the following table
The result of lake classification is shown in figure 3
The area of the lake in the middle stream city group in Yangtze river in the Chinese lake dataset (China lake dataset, CLD) is 8709.56 square kilometers, and the area of the extracted lake (including seasonal wetland and non-seasonal lake) in the experiment is 9512.39 square kilometers. As shown in fig. 3, a and b are two areas densely distributed along the lakes of the Yangtze river, respectively, as can be seen from fig. 3, the lakes with smaller areas in the red dotted line frame are not extracted from the CLD, and the lakes extracted by the method are more complete, and the lakes with smaller areas are also subjected to detailed type recognition
River classification result pairs such as that shown in fig. 4
The river area of the Yangtze river midstream city group in the Chinese marsh wetland space distribution data set (Chinese Academy of Sciences wetland, CAS_Wetlards) is 6620.47 square kilometers, and the extraction result (including seasonal river and non-seasonal river) of the method is 5979.49 square kilometers. As shown in fig. 4, a is a partial enlarged view of the Ganjiang, the cas_wetlands is missing the finer river within the red dotted frame, and the Ganjiang main current water extraction is incomplete. The method has complete detail extraction. b is a partial enlarged view of a river in the Poyang lake flow field, and the CAS_Wetlands is used for extracting an elongated river in a missed manner, so that the classification is more complete by using the method.
Reservoir classification result pairs such as that shown in FIG. 5
The total area of reservoirs in the middle-stream Yangtze city group in the global reservoir and dam database (Global Reservoir and Dam Database, GRAND) is 1849.28 square kilometers, and the area of reservoirs extracted by the research method (comprising quaternary water reservoirs and non-quaternary water reservoirs) is 2584.42 square kilometers. As shown in fig. 5, a and b are mountain areas and hilly areas in the north of the research area, and the GRAND is leaked to extract reservoirs in red dotted frames, so that the extraction result by the method is closer to the real result of manual vectorization, and the type of the water reservoir in the research area is completely identified.
Paddy field and pond classification result pairs are shown in FIG. 6
In the Chinese water body coverage map (China Water Cover Map, CWaC), the rice field area is 13335.65 square kilometers, the pond area is 2060.72 square kilometers, the rice field area is 9468.90 square kilometers, and the pond area is 2091.59 square kilometers. As shown in fig. 6, a is a rice planting area in a Yangtze river midstream city group Dongting lake, the rice extraction result in CWaC is finely divided, the rice field boundary is fuzzy, and the rice field boundary is complete by using the method of the application. b is an enlarged view of the pond in the research area, and compared with a manual vectorization result, the situation that the water body at the edge of the pond and the land part in the CWaC are mixed and separated is shown, and the interval between the ponds is not clear. In the extraction result of the method, the pond water body and the land are distinguished from each other, so that the extraction error is reduced, and the pond identification precision is improved.
The embodiment of the application also provides a water body classification device based on the shape and the water body submerged frequency characteristics, as shown in fig. 7, which is a schematic diagram of a frame structure of the device, and the device comprises:
a first unit 71 for acquiring remote sensing image data;
the method for acquiring remote sensing image data can be realized based on the existing optical satellite, for example, the sentry No. 2 (Sentinel-2) is a sentry series high-resolution optical satellite, and consists of two satellites of 2A and 2B, the revisiting period is 5 days, the revisiting period in a high-latitude area only needs 3 days, the revisiting requirement and the coverage requirement can be met, and the method is widely applied to global higher-resolution earth surface parameter inversion, repeated observation and disaster monitoring research. A multispectral imager (MSI) carried by a Sentinel-2 satellite comprises 13 spectral bands such as visible light, near infrared, short wave infrared and the like, the breadth of an image is 290km, and the spatial resolution is 10m, 20m and 60m. The remote sensing data used in the application is a Sentinel-2 Level 2A product from 1 st 2021 to 31 nd 2021, and the Level-2A image is a standard product of surface reflectivity and is subjected to orthocorrection of the bottom layer reflectivity (BOA) of the atmosphere.
In order to ensure the accuracy of data processing, the embodiment of the application performs preprocessing and correction on acquired remote sensing image data, wherein the preprocessing comprises: wave band combination, cloud cover removal, mosaic, cutting and the like; the correction includes: atmospheric correction, radiation correction, geometric correction and/or resampling, and the like.
A second unit 72, configured to perform a water body index calculation on the remote sensing image data, and segment a water body index calculation result to obtain water body image data;
the specific method for calculating the water body index and dividing the water body index calculation result is the same as that described in the above method embodiment, and will not be described herein again.
A third unit 73 for extracting surface water body shape characteristic information and water body flooding frequency information based on the water body image data;
the application constructs an surface water body classification system based on a water body formation mode and water body inundation frequency information, and mainly comprises the following steps as shown in fig. 2: seasonal wetlands, seasonal rivers, seasonal reservoirs, paddy fields, non-seasonal lakes, non-seasonal rivers, non-seasonal reservoirs, ponds, ditches, and the like.
Different types of surface waters are geometrically different, for example lakes are generally irregular polygons with larger areas, ponds and paddy fields are mostly regular rectangles with smaller areas, rivers are linear water bodies with long and thin curves, reservoirs are complex polygons with curved boundaries, and therefore the water shape characteristics can be selected to primarily distinguish the five categories of water bodies.
It will be appreciated that the body of water shape includes a number of features from which it is desirable to select the body of water shape feature for use in the present invention, and that the third unit 73 of the present embodiment selects the shape feature from a number of surface body of water geometry features by means of a feature importance calculation. For example, 10 commonly used water body Shape features are firstly selected as classification initial features, 5 features are screened from the 10 selected features through feature importance calculation and are combined into a component feature space, and the component feature space is classified according to importance sequences, namely Compactness, area, length, boundary index and Shape index.
The third unit 73 of the present application performs the judgment classification of the permanent water body (0.75. Ltoreq. WIF. Ltoreq.1), the quaternary water body (0.25. Ltoreq. WIF. Ltoreq.0.75) and the maximum water body (0.25. Ltoreq. WIF. Ltoreq.1) by calculating the frequency of occurrence of the water body pixels in the image (i.e., the water body flooding frequency). The water flooding frequency (WIF) refers to the ratio of the number of water floods that occur in a region to the total number of floods that occur in a time series. The water body submerged frequency information is obtained by superposing water body diagrams at different times, and the water body submerged frequency of each pixel in the whole time sequence is calculated by the following steps:
n represents the number of pixel values for all good observations,W ij is a binary variable representing an imagejMiddle pixeli1 for flooding, 0 for non-flooding,O ij representing an imagejMiddle pixeli1 indicates observation, and 0 indicates no observation.
And a fourth unit 74 for classifying the surface water body by using a random forest classifier based on the water body shape characteristic information and the water body inundation frequency information.
According to the water body type constructed by the method, different water bodies of different types have different shape characteristics and water body submerged frequency characteristics, so that the method for constructing the shape characteristics and the water body submerged frequency characteristics is used for carrying out type identification on surface water body refinement.
The shape characteristics of different types of water bodies are different, and firstly, the random forest classifier based on the shape characteristics can realize the identification of five types of water bodies in lakes, rivers, reservoirs, ponds/paddy fields and ditches. Secondly, the non-seasonal lake, the non-seasonal river and the non-seasonal water reservoir are permanent water bodies, the water bodies are covered almost all the year round, the seasonal wetland, the seasonal river and the seasonal water reservoir are seasonal change water bodies, the water bodies are covered only in the water-rich period, and the dead water period is the land, so that the identification of the type can be realized through a random forest classifier based on the water body submerged frequency characteristics.
Therefore, the step of classifying the surface water body by using the random forest classifier based on the water body shape characteristic information and the water body inundation frequency information in the embodiment of the application comprises the following two steps:
1) Performing first-layer classification on the surface water body by using a random forest classifier based on the water body shape characteristic information; the first layer classification is used for distinguishing that the water body category belongs to one of first preset categories, and the first preset categories comprise: lakes, rivers, reservoirs, ponds, paddy fields and ditches.
2) And carrying out second-layer classification on the surface water body by utilizing a random forest classifier based on the water body inundation frequency information. The second layer classification is used for distinguishing that the water body category belongs to one of second preset categories, and the second preset categories comprise: permanent water, seasonal water, and maximum water.
In summary, according to the water body classification method and device disclosed by the embodiment of the invention, remote sensing image data are firstly obtained, image segmentation is carried out on the remote sensing image data to obtain water body image data, and surface water body shape characteristic information and water body inundation frequency information are extracted based on the water body image data, so that surface water bodies are classified by using a random forest classifier based on the water body shape characteristic information and the water body inundation frequency information. The scheme carries out fine classification on the surface water body based on the shape characteristics and the water body inundation frequency characteristics, realizes more accurate water body classification, and provides favorable technical support for water resource management and accurate monitoring and evaluation of ecological environment.
It is to be understood that the invention is not limited to the precise arrangements and instrumentalities shown in the drawings, which have been described above, and that various modifications and changes may be effected without departing from the scope thereof. The scope of the invention is limited only by the appended claims.
The foregoing description of the preferred embodiments of the invention is not intended to limit the invention to the precise form disclosed, and any such modifications, equivalents, and alternatives falling within the spirit and scope of the invention are intended to be included within the scope of the invention.
Claims (10)
1. A water body classification method based on shape and water body submerged frequency features is characterized by comprising the following steps:
acquiring remote sensing image data;
carrying out water body index calculation on the remote sensing image data, and dividing a water body index calculation result to obtain water body image data;
extracting surface water body shape characteristic information and water body inundation frequency information based on the water body image data;
and classifying the surface water body by using a random forest classifier based on the water body shape characteristic information and the water body inundation frequency information.
2. The method of claim 1, wherein the step of segmenting the water body index calculation to obtain water body image data comprises:
threshold segmentation is carried out on the water index calculation result by using an Ojin method, so as to obtain a water-background two-classification chart;
carrying out image gray threshold segmentation on the two classification images, and separating a water body from a background;
and dividing the separated water body by using a multi-scale dividing method to obtain water body image data.
3. The method of claim 1, wherein classifying the surface water body with a random forest classifier based on the water body shape characteristic information and water body flooding frequency information comprises:
performing first-layer classification on the surface water body by using a random forest classifier based on the water body shape characteristic information;
and carrying out second-layer classification on the surface water body by utilizing a random forest classifier based on the water body inundation frequency information.
4. A method according to claim 3, wherein the first layer classification is used to distinguish one of a water body class belonging to a first preset class comprising: lakes, rivers, reservoirs, ponds, paddy fields and ditches.
5. A method according to claim 3, wherein the second layer classification is used to distinguish one of a water body class belonging to a second preset class comprising: permanent water, seasonal water, and maximum water.
6. The method of claim 1, wherein prior to extracting surface water body shape feature information based on the water body image data, the method further comprises:
shape features are selected from a number of surface body geometric features by feature importance calculations.
7. The method of claim 6, wherein the selected shape feature comprises:
compactness, area, length, boundary index, and shape index;
and extracting surface water body shape characteristic information based on the water body image data to obtain the selected shape characteristic information.
8. The method according to claim 1, wherein the water flooding frequency information is obtained by superimposing water graphs of different times, and the water flooding frequency WIF of each pixel in the whole time sequence is calculated by the following method:
,
n represents the number of pixel values for all good observations,W ij is a binary variable representing an imagejMiddle pixeliIs 1 represents flooding, 0 tableThe illustration is not meant to be inundated,O ij representing an imagejMiddle pixeli1 indicates observation, and 0 indicates no observation.
9. The method of claim 1, wherein prior to image segmentation of the remote sensing image data, the method further comprises:
preprocessing and correcting the remote sensing image data.
10. A water body classification device based on shape and water body submerged frequency characteristics, comprising:
the first unit is used for acquiring remote sensing image data;
the second unit is used for carrying out water body index calculation on the remote sensing image data and dividing a water body index calculation result to obtain water body image data;
the third unit is used for extracting surface water body shape characteristic information and water body inundation frequency information based on the water body image data;
and the fourth unit is used for classifying the surface water body by utilizing a random forest classifier based on the water body shape characteristic information and the water body inundation frequency information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311567441.5A CN117611888A (en) | 2023-11-23 | 2023-11-23 | Water body classification method and device based on shape and water body submerged frequency characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311567441.5A CN117611888A (en) | 2023-11-23 | 2023-11-23 | Water body classification method and device based on shape and water body submerged frequency characteristics |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117611888A true CN117611888A (en) | 2024-02-27 |
Family
ID=89947421
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311567441.5A Pending CN117611888A (en) | 2023-11-23 | 2023-11-23 | Water body classification method and device based on shape and water body submerged frequency characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117611888A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118134050A (en) * | 2024-03-23 | 2024-06-04 | 中国科学院地理科学与资源研究所 | Urban flood disaster economic loss prediction method and device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190018918A1 (en) * | 2017-07-11 | 2019-01-17 | The Trustees Of Princeton University | System and method for performing accurate hydrologic determination using disparate weather data sources |
| CN110852262A (en) * | 2019-11-11 | 2020-02-28 | 南京大学 | Agricultural land extraction method based on time sequence top-grade first remote sensing image |
| CN111738144A (en) * | 2020-06-19 | 2020-10-02 | 中国水利水电科学研究院 | Surface water product generation method and system based on Google Earth Engine cloud platform |
| CN116385842A (en) * | 2023-03-17 | 2023-07-04 | 安徽理工大学 | Machine learning water body extraction method integrating multiple features of visible light-infrared-radar images |
-
2023
- 2023-11-23 CN CN202311567441.5A patent/CN117611888A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190018918A1 (en) * | 2017-07-11 | 2019-01-17 | The Trustees Of Princeton University | System and method for performing accurate hydrologic determination using disparate weather data sources |
| CN110852262A (en) * | 2019-11-11 | 2020-02-28 | 南京大学 | Agricultural land extraction method based on time sequence top-grade first remote sensing image |
| CN111738144A (en) * | 2020-06-19 | 2020-10-02 | 中国水利水电科学研究院 | Surface water product generation method and system based on Google Earth Engine cloud platform |
| CN116385842A (en) * | 2023-03-17 | 2023-07-04 | 安徽理工大学 | Machine learning water body extraction method integrating multiple features of visible light-infrared-radar images |
Non-Patent Citations (1)
| Title |
|---|
| 严夏青: "基于多源遥感数据长江干流水体面积时空变化特征的研究", CNKI, 31 January 2023 (2023-01-31), pages 1 - 31 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118134050A (en) * | 2024-03-23 | 2024-06-04 | 中国科学院地理科学与资源研究所 | Urban flood disaster economic loss prediction method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yang et al. | Monthly estimation of the surface water extent in France at a 10-m resolution using Sentinel-2 data | |
| CN111767801B (en) | Remote sensing image water area automatic extraction method and system based on deep learning | |
| Rishikeshan et al. | An automated mathematical morphology driven algorithm for water body extraction from remotely sensed images | |
| Hou et al. | Marine floating raft aquaculture extraction of hyperspectral remote sensing images based decision tree algorithm | |
| CN111738144B (en) | Surface water product generation method and system based on Google Earth Engine cloud platform | |
| Lassalle et al. | Deep learning-based individual tree crown delineation in mangrove forests using very-high-resolution satellite imagery | |
| Chang et al. | Multisensor satellite image fusion and networking for all-weather environmental monitoring | |
| Cavallaro et al. | Automatic attribute profiles | |
| Toosi et al. | Mapping disturbance in mangrove ecosystems: Incorporating landscape metrics and PCA-based spatial analysis | |
| CN109584284B (en) | Hierarchical decision-making coastal wetland ground object sample extraction method | |
| Xiao et al. | Segmentation of multispectral high-resolution satellite imagery using log Gabor filters | |
| CN117611888A (en) | Water body classification method and device based on shape and water body submerged frequency characteristics | |
| Xu et al. | Feature-based constraint deep CNN method for mapping rainfall-induced landslides in remote regions with mountainous terrain: An application to Brazil | |
| CN112418506B (en) | Coastal zone wetland ecological safety pattern optimization method and device based on machine learning | |
| Liu et al. | Mapping China’s offshore mariculture based on dense time-series optical and radar data | |
| CN113887472A (en) | Remote sensing image cloud detection method based on cascade color and texture feature attention | |
| Sekertekin | Potential of global thresholding methods for the identification of surface water resources using Sentinel-2 satellite imagery and normalized difference water index | |
| CN115452759A (en) | River and lake health index evaluation method and system based on satellite remote sensing data | |
| CN117315489B (en) | Water body extraction method and device based on local background characteristic information | |
| Xu et al. | Mapping and analyzing the annual dynamics of tidal flats in the conterminous United States from 1984 to 2020 using Google Earth Engine | |
| Alicandro et al. | Hyperspectral PRISMA and Sentinel-2 Preliminary Assessment Comparison in Alba Fucens and Sinuessa Archaeological Sites (Italy) | |
| Ozdemir et al. | Extraction of Water Bodies from High-Resolution Aerial and Satellite Images Using Visual Foundation Models | |
| He et al. | Development of a tidal flat recognition index based on multispectral images for mapping tidal flats | |
| CN115984689A (en) | Multi-scale earth surface complexity feature extraction and land utilization segmentation method | |
| CN116152655A (en) | Coastal wetland classification method, device, equipment and storage medium |
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 |