CN117409203B - Shallow lake area extraction method - Google Patents
Shallow lake area extraction method Download PDFInfo
- Publication number
- CN117409203B CN117409203B CN202311509181.6A CN202311509181A CN117409203B CN 117409203 B CN117409203 B CN 117409203B CN 202311509181 A CN202311509181 A CN 202311509181A CN 117409203 B CN117409203 B CN 117409203B
- Authority
- CN
- China
- Prior art keywords
- lake
- extracted
- positive
- image
- negative
- 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.)
- Active
Links
- 238000000605 extraction Methods 0.000 title claims abstract description 44
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims abstract description 93
- 238000000034 method Methods 0.000 claims abstract description 37
- 230000011218 segmentation Effects 0.000 claims abstract description 34
- 230000002452 interceptive effect Effects 0.000 claims abstract description 31
- 230000007797 corrosion Effects 0.000 claims abstract description 20
- 238000005260 corrosion Methods 0.000 claims abstract description 20
- 238000012545 processing Methods 0.000 claims abstract description 12
- 238000012549 training Methods 0.000 claims abstract description 11
- 238000011282 treatment Methods 0.000 claims abstract description 11
- 238000012544 monitoring process Methods 0.000 claims description 24
- 238000005520 cutting process Methods 0.000 claims description 9
- 230000000877 morphologic effect Effects 0.000 claims description 3
- 238000013135 deep learning Methods 0.000 abstract description 6
- 238000011835 investigation Methods 0.000 description 9
- 230000000694 effects Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000003993 interaction Effects 0.000 description 3
- 238000003066 decision tree Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000011176 pooling Methods 0.000 description 2
- 230000003321 amplification Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000003139 buffering effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 239000002689 soil Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
- G06T5/30—Erosion or dilatation, e.g. thinning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
-
- 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
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/30—Assessment of water resources
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Medical Informatics (AREA)
- Geometry (AREA)
- Databases & Information Systems (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a shallow lake area extraction method, which comprises the following steps: acquiring sample images and tag data of water bodies and aquatic plants, and performing corrosion and coding treatment on the tag data to obtain positive and negative wave band images; training the interactive semantic segmentation model based on the positive and negative wave band images and the sample image; acquiring a water body index and a vegetation index of a lake to be extracted based on a remote sensing image; respectively carrying out corrosion treatment on the water body index and the vegetation index, and constructing positive and negative wave band images of the lake to be extracted based on the treated data; and processing the remote sensing image and the positive and negative band images of the lake to be extracted based on the trained interactive semantic segmentation model to obtain a lake area extraction result. The shallow lake area generated by the method provided by the invention has higher extraction precision for shallow lakes containing aquatic plants by intelligently setting positive and negative pole wave bands of the interactive deep learning input data, and can realize full-automatic and high-precision lake area extraction.
Description
Technical Field
The invention belongs to the technical fields of remote sensing technology and information science, and particularly relates to a shallow lake area extraction method.
Background
Remote sensing technology is one of the main technologies for dynamically monitoring lake areas in a large scale and in real time. The method for extracting the lake area based on the remote sensing image mainly comprises an index (such as a water index) threshold value, decision tree classification, deep learning and the like. The index threshold method mainly constructs a water index image by utilizing arithmetic operations of wave bands such as medium wave infrared, short wave infrared, red, green and the like according to the wave band characteristics of the remote sensing image, and further sets a threshold value to extract the area of the lake water area. The decision tree classification mainly carries out threshold classification on images in different wavebands, and carries out decision-level fusion on all classification results to realize extraction of lake areas, the extraction precision of the method is directly related to each threshold classification result, decision-level fusion rule and the like, the method generally has no generalization capability, an extraction model of one lake area can not be used for other lakes basically, and the engineering capability is poor. The deep learning method realizes extraction of lake areas by training semantic segmentation deep learning network parameters through a large number of priori lake samples and processing lake area images, and the method is limited by factors such as sample quantity quality, type comprehensiveness, distribution uniformity and the like, and needs to consume a large amount of manpower to manufacture samples and time to iterate the samples, so that engineering cost is high.
In addition, as a large number of aquatic plants are often associated in a plurality of shallow lakes, the area covered by the aquatic plants in a region which is formed by the aquatic plants in a slice and is higher than the water surface is often expressed as a vegetation type on a remote sensing image, the three methods can not successfully extract the area of the covered region of the aquatic plants basically, the accuracy of lake area extraction is greatly affected, and the accuracy of lake area data is obviously reduced.
Disclosure of Invention
The invention aims to provide a shallow lake area extraction method for solving the problems in the prior art.
In order to achieve the above object, the present invention provides a method for shallow lake area extraction, comprising the steps of:
acquiring sample images and tag data of water bodies and aquatic plants, and performing corrosion and coding treatment on the tag data to obtain positive and negative wave band images;
training the interactive semantic segmentation model based on the positive and negative wave band images and the sample image to obtain a trained interactive semantic segmentation model;
acquiring a water body index and a vegetation index of a lake to be extracted based on a remote sensing image;
respectively carrying out corrosion treatment on the water body index and the vegetation index, and constructing positive and negative wave band images of the lake to be extracted based on the treated data;
and processing the remote sensing image and the positive and negative band images of the lake to be extracted based on the trained interactive semantic segmentation model to obtain a lake area extraction result.
Optionally, the process of acquiring sample images and tag data of the body of water and the aquatic plant includes: acquiring relevant monitoring data of a water body and aquatic plants and corresponding remote sensing images, and cutting the remote sensing images of the same spatial position area based on the relevant monitoring data of the water body and the aquatic plants to obtain image blocks of image spots of the water body and the aquatic plants; and performing interactive segmentation processing on the water body and the image blocks of the pattern spot area of the aquatic plant based on the pre-training weight of the interactive semantic segmentation model to obtain sample images and label data of the water body and the aquatic plant.
Optionally, the process of corroding the tag data includes: and carrying out omnidirectional recursive corrosion operation on the label data based on the grid label to obtain positive and negative points of the label data.
Optionally, the process of encoding the tag data includes: and converting positive and negative points of the tag data into positive and negative wave band images based on a weighted distance coding algorithm.
Optionally, the process of obtaining the water body index and the vegetation index of the lake to be extracted includes: obtaining a maximum range of the water surface area of a lake to be extracted and a corresponding remote sensing image, and cutting the remote sensing image of the same space position area based on the maximum range of the water surface area of the lake to be extracted to obtain an image block of the image spot area of the lake to be extracted; and acquiring a water body index and a vegetation index of the lake image to be extracted based on the lake image spot region image block to be extracted.
Optionally, the corrosion treatment of the water body index and the vegetation index respectively comprises: carrying out omnidirectional corrosion operation on the water index in the lake to be extracted, and extracting the water seed points in the lake to be extracted; carrying out directional corrosion operation on the vegetation indexes in the lake to be extracted, wherein the directional direction is the same direction as the nearest water body seed point, and extracting the vegetation seed points in the lake to be extracted; carrying out directional morphological operation on the vegetation index outside the lake to be extracted, wherein the directional direction is the opposite direction of the minimum value of the nearest water body seed point and the nearest vegetation seed point, and extracting the vegetation seed point outside the lake to be extracted.
Optionally, the process of constructing the positive and negative band images of the lake to be extracted includes: equally dividing the image blocks of the lake pattern area to be extracted into 3*3 areas; selecting a water seed point and a vegetation seed point which are relatively close to the central point of the lake to be extracted in each area as positive points; selecting a vegetation seed point relatively close to the central point of the lake to be extracted from outside the lake to be extracted in each area as a negative point; and constructing positive and negative wave band images of the lake to be extracted based on the positive points and the negative points.
Optionally, the process of obtaining the lake area extraction result includes: performing interactive semantic segmentation processing on the positive and negative wave band images of the lake to be extracted and the combined image of the lake image to be extracted based on the trained interactive semantic segmentation model; binarization and vectorization are carried out on the treatment result to obtain a primary lake area extraction result; vector cutting is carried out on the initial lake area extraction result by adopting the maximum range of the lake water surface area to be extracted, and the lake area extraction result is obtained.
The invention has the technical effects that:
the shallow lake area generated by the method does not need to set a threshold value manually and does not need to interpret expertise remotely, the extraction precision of the shallow lake containing aquatic plants is higher by intelligently setting positive and negative pole wave bands of the interactive deep learning input data, and full-automatic and high-precision lake area extraction can be realized.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application, illustrate and explain the application and are not to be construed as limiting the application. In the drawings:
FIG. 1 is a schematic flow chart of a method for shallow lake area extraction in an embodiment of the invention;
FIG. 2 is a schematic diagram of an image block of a lake plaque area to be extracted according to an embodiment of the present invention;
FIG. 3 shows the effect of shallow lake area extraction in an embodiment of the present invention; wherein, (a) is a lake region remote sensing image (R: 4, G:3, B: 2), (b) is a water body index image, and (c) is a vegetation index image; (d) The method comprises the steps of (1) taking a seed point image of the vegetation in the lake to be extracted, (e) taking a seed point image of the vegetation outside the lake to be extracted, (f) taking a seed point image of the water in the lake to be extracted, (g) taking positive and negative point images of the lake to be extracted, and (h) taking an area extraction result image of the shallow lake.
Detailed Description
It should be noted that, in the case of no conflict, the embodiments and features in the embodiments may be combined with each other. The present application will be described in detail below with reference to the accompanying drawings in conjunction with embodiments.
It should be noted that the steps illustrated in the flowcharts of the figures may be performed in a computer system such as a set of computer executable instructions, and that although a logical order is illustrated in the flowcharts, in some cases the steps illustrated or described may be performed in an order other than that illustrated herein.
Example 1
As shown in fig. 1, the embodiment provides a method for extracting a shallow lake area, where the method for extracting a lake area includes: utilizing natural resource investigation monitoring data, and based on an interactive semantic segmentation network and pre-training weight parameters, making sample images and tag data of water bodies and aquatic plants; constructing positive, negative point and positive and negative wave band images by using the omnidirectional corrosion and the weighted distance coding and using the positive and negative point and the positive and negative wave band images together with the sample images for updating parameters of the interactive semantic segmentation model to obtain the interactive semantic segmentation model; obtaining the maximum range of the shallow lake area to be extracted according to the natural resource investigation monitoring data; calculating a water body index and a vegetation index of an image of a shallow lake area to be extracted by using remote sensing images including mid-infrared, red and green wave bands such as Landsat7/8, and respectively extracting water body and vegetation seed points inside and outside the shallow lake area through omnidirectional and directional corrosion operation; constructing positive and negative band images by taking water seed points and vegetation seed points inside and outside a lake area as positive and negative points; and carrying out interactive semantic segmentation correction on the pixels with wrong extraction results by using an interactive semantic segmentation model to obtain lake area extraction results.
The specific implementation method comprises the following steps:
and collecting third national soil investigation data, geographical national condition monitoring data, a southwest island land area water body remote sensing extraction data set, a national earth surface type remote sensing image sample data set and other natural resource investigation monitoring data, and processing to form water body related monitoring data.
The natural resource survey monitoring data can be expressed as:
R＝{P i ,i＝1,2,....,N}
wherein R is a natural resource investigation monitoring data set, P i For each type of survey monitoring data, N is the total number of data.
Monitoring data P for each class of surveys i Combining according to the pattern spot type attribute to obtain relevant data P of each type of investigation monitoring water body and aquatic plant i ’。
P i ’＝{UNION Pij J e field in (DLBM, DLDM, LXBH, code of water, aquatic plant related categories, or DLMC, lbmcHydraulic facilities, lakes, rivers, aquatic plants, etc.) }.
Wherein P is i ' monitoring data related to water and aquatic plants for each type of investigation, UNION is UNION, P ij Is P i Field is an attribute field, (DLBM, DLDM, LXBH,) is a category code attribute column, (DLMC, LBMC,) is a category name attribute column, or is or.
For all survey monitoring data sets R, P i 'merging according to the space position to obtain relevant monitoring data R' of the water body and the aquatic plants:
R’＝{UNION Pi’ }
wherein R' is water body and aquatic plant related monitoring data, P i ' is survey monitoring data of each type after category combination, and UNION is UNION.
And cutting the optical remote sensing image of the same spatial position area by using the water body and aquatic plant related monitoring data R', and carrying out interactive segmentation processing on the cut image blocks based on the lightweight interactive semantic segmentation network pre-training weight to manufacture sample images and label data of the shallow water surface.
As a specific embodiment, as shown in fig. 2, each image spot minimum outer-section rectangle of the water body and aquatic plant related monitoring data R' is buffered outwards, and the buffered rectangle is used for cutting the optical remote sensing image of the same space position area to obtain the image block of the water body and aquatic plant image spot area. The method is characterized in that the method comprises the steps of respectively buffering the image according to the equal proportion of 30% of the length and width of the minimum outer section rectangle of the image spots to the outside, wherein the length and width of the buffered rectangle are 160% of the original image, the image is subjected to repeated pooling to obtain context semantic information of a large step length of the image in consideration of a semantic segmentation model, the size of the image is 1/5-1/10 of the size of the original image after 3-4 times of pooling, and the outermost layer pixel can contain background information outside the water body so as to be beneficial to balancing positive and negative samples of the water body and aquatic plants.
Based on the HRNet18s+OCR lightweight interactive semantic segmentation network model, a remote sensing image is referenced, a visual interpretation mode is adopted, a water surface area pattern spot containing aquatic plants is drawn on the remote sensing image, and sample image and label data of a water body are obtained. And carrying out amplification treatments such as translation, rotation, overturning, scaling, noise increasing, brightness stretching, color conversion, splicing, filling, stacking and the like on the sample data according to the proportion of not more than 50%, and sampling and undersampling to form water sample data with balanced category distribution.
sample＝{Sample water or Sample hydrophyte }
Wherein Sample is a water Sample data set water Sample is water surface Sample data hydrophyte For aquatic plant sample data, or is or.
And performing grid-controlled corrosion operation on the tag data, constructing positive and negative point band images, and inputting HRNet18s+ OCR interaction semantic segmentation network together with the sample image for training to obtain an interaction semantic segmentation model to obtain an interaction semantic segmentation network model.
As a specific embodiment, performing omnidirectional recursive corrosion operation of positive and negative pixels on the tag data according to the grid to obtain positive and negative points of the tag data.
Point_positive＝{Q ij ∈(Sample label ＝1)，i＝[1,3],j＝[1,3]}
Point_nagetive＝{Q ij ∈(Sample label ＝0)，i＝[1,3],j＝[1,3]}
B i ＝1{i＝1,2,3,4,5,6,7,8}
Wherein, point_positive is positive pixel, point_negative is negative pixel, sample label Tag image of water body sample, Q ij For the grid label image, B is the 3*3 convolution kernel.
And converting the positive and negative points of the label data into positive and negative band images by adopting a weighted distance coding algorithm. All positive points and all negative points are respectively encoded.
D ij ＝sqrt((x i -x j )^2+(y i -y j )^2)
D max ＝MAX(D ij )
Point_mid＝(Σx i /N,Σy i /N)i＝1,...,N
Gray k ＝(D max -D point_mid,k )
Wherein D is ij Point_mid is the center Point of all positive or negative points, D, for the distance between any two points max Is the maximum value of the distance between any two points, D point_mid,k For the distance between pixel position k and the center Point point_mid, N is the total number of positive/negative points, gray k Is the gray value of position k in the positive or negative point band image.
And carrying out interactive semantic segmentation model updating training on the water body sample data, the tag data and the positive and negative point wave band images by using HRNet18s+OCR interactive semantic segmentation pre-training parameters.
And obtaining the maximum range of the water surface area of the shallow lake by using natural resource investigation monitoring data, and processing remote sensing images including mid-infrared, red and green wave bands such as Landsat7/8 by using an HRNet18s+OCR interactive semantic segmentation model to extract the water area of the shallow lake.
Obtaining the maximum range of the water surface area of the shallow lake according to the natural resource investigation and monitoring data;
P k ’＝{UNION Pij j e field in (DLMC, lbmc. = shallow water lake name, etc.) }
Obtaining shallow lake area images to be extracted in the same way as the step 2 for remote sensing images including middle infrared, red and green wave bands such as Landsat7/8 and calculating the water body index and vegetation index of the images to be extracted;
according to the maximum range of the water surface area of the shallow lake, carrying out omnidirectional corrosion operation on the water index in the lake area, and extracting water seed points of the image to be extracted;
B＝B i ＝1{i＝1,2,3,4,5,6,7,8}
wherein, NDWI in The water index image in the lake area is shown, and B is 3*3 convolution kernel.
Carrying out directional corrosion operation on vegetation indexes in a lake area according to the maximum range of the water surface area of the shallow lake, wherein the directional direction is the same direction of the nearest water seed point, and extracting vegetation seed points in the lake area;
B＝{B i =1 if i in direction of nearest water extreme point, bi=0 else }
Wherein, NDVI in The vegetation index image in the lake region is shown, and B is 3*3 convolution kernel.
Carrying out directional morphological operation on the vegetation indexes outside the lake area, wherein the directional direction is the opposite direction of the minimum value of the nearest water body seed point and the nearest vegetation seed point, and extracting the vegetation seed points outside the lake area;
b= { bi=1 if i in reverse of the minimum of the nearest water/vegetation seed point, bi=0 else }
Wherein NDVIout is a vegetation index image outside the lake area, and B is a 3*3 convolution kernel.
Lake area extraction:
equally dividing the region to be extracted into 3*3 regions; respectively selecting a water seed point and a vegetation seed point which are relatively close to the central point of the lake area as positive points for the water seed point and the vegetation seed point in the lake area; and respectively selecting a vegetation seed point which is relatively close to the central point of the lake area as a negative point for the vegetation seed points outside the lake area.
Constructing a positive point distance wave band image and a negative point distance wave band image by positive points and negative points; and performing interactive semantic segmentation processing on the combined images such as the positive point wave band image, the negative point wave band image, the original image and the like by using the trained interactive semantic segmentation network model.
Binarization and vectorization are carried out on the deep learning result to obtain a primary lake area extraction result; vector cutting is carried out on the initial lake area extraction result by adopting the lake area range, and the lake area extraction result is obtained. As shown in fig. 3, the effect of shallow lake area extraction in the embodiment of the invention is shown; wherein, (a) is a lake region remote sensing image (R: 4, G:3, B: 2), (b) is a water body index image, and (c) is a vegetation index image; (d) The method comprises the steps of (1) taking a seed point image of the vegetation in the lake to be extracted, (e) taking a seed point image of the vegetation outside the lake to be extracted, (f) taking a seed point image of the water in the lake to be extracted, (g) taking positive and negative point images of the lake to be extracted, and (h) taking an area extraction result image of the shallow lake.
The foregoing is merely a preferred embodiment of the present application, but the scope of the present application is not limited thereto, and any changes or substitutions easily contemplated by those skilled in the art within the technical scope of the present application should be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (6)
1. A method for shallow lake area extraction, comprising the steps of:
acquiring sample images and tag data of water bodies and aquatic plants, and performing corrosion and coding treatment on the tag data to obtain positive and negative wave band images;
training the interactive semantic segmentation model based on the positive and negative wave band images and the sample image to obtain a trained interactive semantic segmentation model;
acquiring a water body index and a vegetation index of a lake to be extracted based on a remote sensing image;
respectively carrying out corrosion treatment on the water body index and the vegetation index, and constructing positive and negative wave band images of the lake to be extracted based on the treated data;
processing the remote sensing image and the positive and negative band images of the lake to be extracted based on the trained interactive semantic segmentation model to obtain a lake area extraction result;
the corrosion treatment process for the water body index and the vegetation index respectively comprises the following steps: carrying out omnidirectional corrosion operation on the water index in the lake to be extracted, and extracting the water seed points in the lake to be extracted; carrying out directional corrosion operation on the vegetation indexes in the lake to be extracted, wherein the directional direction is the same direction as the nearest water body seed point, and extracting the vegetation seed points in the lake to be extracted; carrying out directional morphological operation on the vegetation indexes outside the lake to be extracted, wherein the directional direction is the opposite direction of the minimum value of the nearest water body seed point and the nearest vegetation seed point, and extracting the vegetation seed points outside the lake to be extracted;
the process for constructing the positive and negative wave band images of the lake to be extracted comprises the following steps: equally dividing the image blocks of the lake pattern area to be extracted into 3*3 areas; selecting a water seed point and a vegetation seed point which are relatively close to the central point of the lake to be extracted in each area as positive points; selecting a vegetation seed point relatively close to the central point of the lake to be extracted from outside the lake to be extracted in each area as a negative point; and constructing positive and negative wave band images of the lake to be extracted based on the positive points and the negative points.
2. The method for shallow lake area extraction according to claim 1, wherein,
the process of acquiring sample images and tag data of water and aquatic plants includes: acquiring relevant monitoring data of a water body and aquatic plants and corresponding remote sensing images, and cutting the remote sensing images of the same spatial position area based on the relevant monitoring data of the water body and the aquatic plants to obtain image blocks of image spots of the water body and the aquatic plants; and performing interactive segmentation processing on the water body and the image blocks of the pattern spot area of the aquatic plant based on the pre-training weight of the interactive semantic segmentation model to obtain sample images and label data of the water body and the aquatic plant.
3. The method for shallow lake area extraction according to claim 1, wherein,
the process of corroding the tag data comprises the following steps: and carrying out omnidirectional recursive corrosion operation on the label data based on the grid label to obtain positive and negative points of the label data.
4. A method for shallow lake area extraction according to claim 3, wherein,
the process of encoding the tag data includes: and converting positive and negative points of the tag data into positive and negative wave band images based on a weighted distance coding algorithm.
5. The method for shallow lake area extraction according to claim 1, wherein,
the process for obtaining the water body index and the vegetation index of the lake to be extracted comprises the following steps: obtaining a maximum range of the water surface area of the lake to be extracted and a corresponding remote sensing image, cutting the remote sensing image of the same space position area based on the maximum range of the water surface area of the lake to be extracted, obtaining an image block of the image spot area of the lake to be extracted, and obtaining the water body index and the vegetation index of the lake image to be extracted based on the image block of the image spot area of the lake to be extracted.
6. The method for shallow lake area extraction according to claim 1, wherein,
the process for obtaining the lake area extraction result comprises the following steps: performing interactive semantic segmentation processing on the positive and negative wave band images of the lake to be extracted and the combined image of the lake image to be extracted based on the trained interactive semantic segmentation model; binarization and vectorization are carried out on the treatment result to obtain a primary lake area extraction result; vector cutting is carried out on the initial lake area extraction result by adopting the maximum range of the lake water surface area to be extracted, and the lake area extraction result is obtained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311509181.6A CN117409203B (en) | 2023-11-14 | 2023-11-14 | Shallow lake area extraction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311509181.6A CN117409203B (en) | 2023-11-14 | 2023-11-14 | Shallow lake area extraction method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117409203A CN117409203A (en) | 2024-01-16 |
| CN117409203B true CN117409203B (en) | 2024-04-02 |
Family
ID=89494373
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311509181.6A Active CN117409203B (en) | 2023-11-14 | 2023-11-14 | Shallow lake area extraction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117409203B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104915954A (en) * | 2015-05-25 | 2015-09-16 | 同济大学 | Automatic water body extraction method based on Landsat OLI multispectral remote sensing image |
| CN108734122A (en) * | 2018-05-17 | 2018-11-02 | 北京理工大学 | A kind of EO-1 hyperion city water body detection method based on adaptive samples selection |
| CN111767801A (en) * | 2020-06-03 | 2020-10-13 | 中国地质大学（武汉） | Remote sensing image water area automatic extraction method and system based on deep learning |
| CN111931696A (en) * | 2020-09-04 | 2020-11-13 | 中国水利水电科学研究院 | Lake and reservoir water area remote sensing automatic extraction method based on space constraint |
| CN114332619A (en) * | 2021-12-30 | 2022-04-12 | 安徽大学 | Remote sensing data bloom extraction method based on deep learning |
| WO2022077958A1 (en) * | 2020-10-13 | 2022-04-21 | 上海商汤智能科技有限公司 | Water-body segmentation method and apparatus, electronic device, and storage medium |
| CN114419460A (en) * | 2022-01-19 | 2022-04-29 | 中国科学院空天信息创新研究院 | Water body area extraction method and device |
| CN115410091A (en) * | 2022-07-28 | 2022-11-29 | 山东电力工程咨询院有限公司 | Method and system for extracting lake water area of arid region |
-
2023
- 2023-11-14 CN CN202311509181.6A patent/CN117409203B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104915954A (en) * | 2015-05-25 | 2015-09-16 | 同济大学 | Automatic water body extraction method based on Landsat OLI multispectral remote sensing image |
| CN108734122A (en) * | 2018-05-17 | 2018-11-02 | 北京理工大学 | A kind of EO-1 hyperion city water body detection method based on adaptive samples selection |
| CN111767801A (en) * | 2020-06-03 | 2020-10-13 | 中国地质大学（武汉） | Remote sensing image water area automatic extraction method and system based on deep learning |
| CN111931696A (en) * | 2020-09-04 | 2020-11-13 | 中国水利水电科学研究院 | Lake and reservoir water area remote sensing automatic extraction method based on space constraint |
| WO2022077958A1 (en) * | 2020-10-13 | 2022-04-21 | 上海商汤智能科技有限公司 | Water-body segmentation method and apparatus, electronic device, and storage medium |
| CN114332619A (en) * | 2021-12-30 | 2022-04-12 | 安徽大学 | Remote sensing data bloom extraction method based on deep learning |
| CN114419460A (en) * | 2022-01-19 | 2022-04-29 | 中国科学院空天信息创新研究院 | Water body area extraction method and device |
| CN115410091A (en) * | 2022-07-28 | 2022-11-29 | 山东电力工程咨询院有限公司 | Method and system for extracting lake water area of arid region |
Non-Patent Citations (1)
| Title |
|---|
| 基于知识决策树的城市水体提取方法研究;陈静波 等;遥感信息;20130228;第28卷(第1期);29-33 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117409203A (en) | 2024-01-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111767801B (en) | Remote sensing image water area automatic extraction method and system based on deep learning | |
| CN113011427B (en) | Remote sensing image semantic segmentation method based on self-supervision contrast learning | |
| Alhassan et al. | A deep learning framework for land-use/land-cover mapping and analysis using multispectral satellite imagery | |
| CN112991354A (en) | High-resolution remote sensing image semantic segmentation method based on deep learning | |
| CN113312993B (en) | Remote sensing data land cover classification method based on PSPNet | |
| CN112561876A (en) | Image-based pond and reservoir water quality detection method and system | |
| CN112950780A (en) | Intelligent network map generation method and system based on remote sensing image | |
| Zhang et al. | Segmentation model based on convolutional neural networks for extracting vegetation from Gaofen-2 images | |
| CN114494910B (en) | Multi-category identification and classification method for facility agricultural land based on remote sensing image | |
| CN117058367A (en) | Semantic segmentation method and device for high-resolution remote sensing image building | |
| CN117975282B (en) | Remote sensing extraction method and system for laver cultivation area based on multi-element optical feature fusion | |
| Mantripragada et al. | The effects of spectral dimensionality reduction on hyperspectral pixel classification: A case study | |
| Gonzalez | DL4DS—Deep learning for empirical downscaling | |
| Yuan et al. | MPFFNet: LULC classification model for high-resolution remote sensing images with multi-path feature fusion | |
| Li et al. | Fruit tree canopy segmentation from UAV orthophoto maps based on a lightweight improved U-Net | |
| CN117274834A (en) | Water remote sensing extraction method considering optical characteristic multi-scale fusion | |
| CN111832480B (en) | Remote sensing identification method for rape planting area based on spectral characteristics | |
| CN116543165B (en) | Remote sensing image fruit tree segmentation method based on dual-channel composite depth network | |
| CN117611888A (en) | Water body classification method and device based on shape and water body submerged frequency characteristics | |
| CN117409203B (en) | Shallow lake area extraction method | |
| CN117455939A (en) | Cultivated land boundary extraction method of remote sensing satellite image | |
| CN117893744A (en) | Remote sensing image segmentation method based on improved boundary guide context aggregation network | |
| CN116385875A (en) | Remote sensing vegetation extraction method and device based on deep learning semantic segmentation | |
| Yi et al. | AC-UNet: an improved UNet-based method for stem and leaf segmentation in Betula luminifera | |
| CN113642663B (en) | Satellite remote sensing image water body extraction method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |