CN113705523B - Layered urban impervious surface extraction method based on optical and dual-polarized SAR fusion - Google Patents
Layered urban impervious surface extraction method based on optical and dual-polarized SAR fusion Download PDFInfo
- Publication number
- CN113705523B CN113705523B CN202111037677.9A CN202111037677A CN113705523B CN 113705523 B CN113705523 B CN 113705523B CN 202111037677 A CN202111037677 A CN 202111037677A CN 113705523 B CN113705523 B CN 113705523B
- Authority
- CN
- China
- Prior art keywords
- dual
- shadow
- optical
- polarized sar
- polarized
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
Abstract
The invention discloses a layered urban impervious surface extraction method based on optical and dual-polarized SAR fusion, which is used for respectively preprocessing optical and dual-polarized SAR images; carrying out strict geographic registration on the preprocessed optical image and the dual-polarized SAR image; calculating spectral features and texture features of the preprocessed optical image; and calculating polarization characteristics and texture characteristics of the preprocessed dual-polarized SAR data. The invention provides a layered urban impervious surface extraction method based on the fusion of optical and dual-polarized SAR, which fully utilizes the information of two data sources of the optical and dual-polarized SAR, improves the urban impervious surface extraction precision, fully considers the imaging principle and the data characteristics of the optical and dual-polarized SAR image, extracts the characteristic with stronger distinguishing property to the impervious surface, particularly the polarization characteristic, and adopts an upper layered classification frame, thereby simultaneously reducing the shadow problem of the urban center and the bare soil problem of the urban edge and providing an effective solution for the urban impervious surface extraction with a larger range.
Description
Technical Field
The invention relates to the field of impervious surface extraction, in particular to a layered urban impervious surface extraction method based on optical and dual-polarized SAR fusion.
Background
At present, the global urbanization process is rapid, and the global urbanization rate exceeds 50% in 2018. The urbanization process is accompanied by a large number of artificial impermeable surfaces instead of natural ground surfaces, so that the impermeable surfaces represent an important index of the urbanization standard. The great increase of the impermeable surface brings great pressure to urban management and ecological environment, so that the extraction of the impermeable surface of the city has important significance.
At present, urban impervious surface extraction faces two important challenges, and a large number of shadows caused by forestation of high buildings in the city cause dark impervious surfaces, so that the shadows are confused with water bodies; the urban expansion causes that a large amount of bare soil at the urban edge is confused with a bright impervious surface, and the extraction precision of the impervious surface is seriously reduced. Therefore, the accuracy of extracting the water impermeable surface by only using the optical image is limited. The SAR data can supplement the optical data due to different imaging modes, and the extraction precision of the impermeable surface can be improved by combining the optical data with the SAR data. However, currently, there are few fusion experiments using dual polarized SAR and optical images. The dual polarized SAR can provide sufficient polarization information while the amount of data is smaller. However, the polarization characteristics of the water impermeable surface extracted by the dual-polarized SAR are still in a blank stage, so that the polarization information of the dual-polarized SAR cannot be fully utilized due to the lack of the polarization characteristics, and the difficult problem faced by the urban water impermeable surface extraction cannot be effectively solved. Therefore, we propose a layered urban impervious surface extraction method based on the fusion of optical and dual-polarized SAR.
Disclosure of Invention
Based on the technical problems in the background art, the invention provides a layered urban impervious surface extraction method based on the fusion of optical and dual-polarized SAR, which has the characteristics of improving the impervious surface extraction precision and solving the problem that the urban impervious surface extraction is difficult because a large amount of shadows in the center of the city and a large amount of bare soil at the edge of the city are similar to the impervious surface spectrum.
The invention provides the following technical scheme:
a layered urban impervious surface extraction method based on optical and dual-polarization SAR fusion comprises the following steps:
A. the image feature extraction based on the characteristic of the water-impermeable surface specifically comprises the following steps:
a1, respectively preprocessing the optical and dual-polarized SAR images;
a2, carrying out strict geographic registration on the preprocessed optical image and the dual-polarized SAR image;
a3, calculating spectral characteristics and texture characteristics of the preprocessed optical image;
a4, calculating polarization characteristics and texture characteristics of the preprocessed dual-polarized SAR data;
B. urban impervious surface extraction based on layered framework, and the specific process comprises the following steps:
b.1, primarily classifying the images to obtain non-shadow vegetation, non-shadow impermeable water, non-shadow water body, non-shadow bare soil and shadow;
b.2, selecting characteristics, and reclassifying the shadow obtained in the step S2.1 to obtain shadow vegetation, a shadow water body, shadow bare soil and a shadow impermeable water body;
and B.3, combining the non-shadow impermeable surface obtained in the step S2.1 and the shadow impermeable surface obtained in the step S2.2 to obtain impermeable surface information in the whole area.
Preferably, in the step a.3, a normalized vegetation index NDVI and a normalized water body index are calculated for the preprocessed optical image; and then respectively adopting gray level co-occurrence matrixes to calculate texture characteristics for visible light wave bands including a blue light wave band, a green light wave band, a red light wave band and a near infrared wave band.
Preferably, in the step a.4, the polarization characteristics of the preprocessed dual-polarized SAR data are obtained through a covariance matrix, and the texture characteristics are calculated by respectively adopting a gray level co-occurrence matrix for the planned image.
Preferably, all optical and dual polarized SAR features are stacked in b.1, and initially classified using random forests.
Preferably, the dual polarized SAR characteristics and NDVI stack are selected in the b.2 to be input into a random forest, and the shadows obtained in the step b.1 are reclassified.
The invention provides a layered urban impervious surface extraction method based on optical and dual-polarized SAR fusion, which selects more effective dual-polarized SAR polarization characteristics for extracting impervious surfaces according to an image principle by taking urban optical remote sensing images and dual-polarized SAR images as main data sources, improves the problem of confusion between bare soil and bright impervious surfaces, aims at urban shadow problems, designs a layered classification frame, avoids classifying shadows directly, fully utilizes information of the optical and dual-polarized SAR data sources, improves urban impervious surface extraction precision, fully considers imaging principles and data characteristics of the optical and dual-polarized SAR images, extracts characteristics with stronger differentiation on the impervious surfaces, particularly polarization characteristics, and improves the layered classification frame, can simultaneously reduce the shadow problem of urban centers and the bare soil problem of urban edges, and provides an effective solution for urban impervious surface extraction in a larger range.
Drawings
FIG. 1 is a flow chart of a layered urban impervious surface extraction method based on the fusion of optical and dual-polarized SAR;
FIG. 2 is a dual polarized SAR polarization feature gray scale histogram;
FIG. 3 is a dual polarized SAR polarization profile;
fig. 4 is a graph of the result of extraction of the impervious surface, wherein fig. 4a: the impervious surface extraction map obtained by the method is shown in the specification; fig. 4b: the resulting impervious surface extraction map is simply stacked.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. 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.
Referring to fig. 1-4, the present invention provides a technical solution:
a layered urban impervious surface extraction method based on optical and dual-polarization SAR fusion comprises the following steps:
A. the image feature extraction based on the characteristic of the water-impermeable surface specifically comprises the following steps:
a1, respectively preprocessing the optical and dual-polarized SAR images;
a2, carrying out strict geographic registration on the preprocessed optical image and the dual-polarized SAR image;
a3, calculating spectral characteristics and texture characteristics of the preprocessed optical image;
a4, calculating polarization characteristics and texture characteristics of the preprocessed dual-polarized SAR data;
B. urban impervious surface extraction based on layered framework, and the specific process comprises the following steps:
b.1, primarily classifying the images to obtain non-shadow vegetation, non-shadow impermeable water, non-shadow water body, non-shadow bare soil and shadow;
b.2, selecting characteristics, and reclassifying the shadow obtained in the step S2.1 to obtain shadow vegetation, a shadow water body, shadow bare soil and a shadow impermeable water body;
and B.3, combining the non-shadow impermeable surface obtained in the step S2.1 and the shadow impermeable surface obtained in the step S2.2 to obtain impermeable surface information in the whole area.
Example 1
A. Image feature extraction based on water impermeable surface characteristics
Because of different imaging mechanisms, the impermeable surface has different characteristics in the optical image and the dual-polarized SAR image, the characteristics can be complemented, and the extraction precision of the impermeable surface is improved together, and the method specifically comprises the following steps:
and A.1, respectively preprocessing the optical and dual-polarized SAR images:
pretreatment of the optical image (mainly for middle-high resolution images like Sentinel-2) includes radiation correction, atmospheric correction; preprocessing dual polarized SAR images, including amplitude separation (split), de-banding (de-burst), radiation correction (radiometric calibrate), speckle filtering (speckle filter) and terrain correction (terrain correction); in the example, SNAP software is adopted for preprocessing the optical and dual-polarized SAR images;
and A.2, carrying out strict geographic registration on the preprocessed optical image and the dual-polarized SAR image:
for a multisource data fusion experiment, the registration of influence data of different sources is very important, the subsequent fusion result is directly influenced by the registration precision, and in the example, the EVNI software is used for registration;
and A.3, calculating spectral characteristics and texture characteristics of the preprocessed optical image:
to increase the distinguishing of the impermeable surface and the vegetation and water, respectively calculating NDVI and NDWI, the texture information can provide important space and structure information in the urban classification, so that a gray level co-occurrence matrix is adopted to respectively calculate the median (mean), the covariance (covariance), the contrast (contrast) and the dissimilarity (dissimilarity) of the visible light bands R, G, B and Nir of the optical image;
and A.4, calculating polarization characteristics and texture characteristics of the preprocessed dual-polarization SAR data:
the polarization characteristics are selected from diagonal elements of covariance matrix C of the preprocessed dual-polarized SAR image, and C is defined as follows:
wherein S is VV Is a scattering matrix of VV polarization, S VH Is a VH polarized scattering matrix, x is defined as:
x＝[S VV S VH ] T
texture feature selection median (mean), homogeneity (homogeneity) and covariance (variance), each polarized image is computed separately, in this example, extracted polarization feature C 11 And C t2 The gray level histograms of (a) and (b) in fig. 2 are shown respectively, wherein, there are several obvious peaks, the peak closest to the origin is known to represent the water body by the data characteristics of the SAR, and the peak farthest from the origin is referred to as the impermeable water surface, thus, the difference between the water body and the shadow is increased, the characteristic diagram is shown in (a) and (b) in fig. 3, and it can be seen that, in the polarization characteristics extracted by the invention, the brightness of the impermeable water surface is enhanced, and the intra-class difference from the permeable water surface is increased;
B. urban impervious surface extraction based on layered framework:
the existing optical and SAR fusion method is characterized in that all the characteristics of the optical and SAR are simply stacked, then the optical and SAR are used as input of a classifier to obtain a classification result of the impermeable surface, shadows cannot be effectively identified by the method, and therefore extraction accuracy of the impermeable surface is reduced;
b.1, primarily classifying the images to obtain non-shadow vegetation, non-shadow impermeable water, non-shadow water body, non-shadow bare soil and shadow;
in the step, all the features extracted from the optical and SAR images are classified as random forest input so as to ensure the optimal classification precision and obtain accurate non-shadow impermeable water surface and shadow;
b.2, selecting characteristics, and reclassifying the shadow obtained in the step S2.1 to obtain shadow vegetation, a shadow water body, shadow bare soil and a shadow impermeable water body;
wherein, because the spectrum information in the shadow is weak, only SAR related features are selected; in addition, when SAR polarization characteristics are used, confusion between vegetation and impermeable water is not negligible, and the NDVI value of the vegetation is higher and the trend is still reserved in shadows, so that the NDVI and SAR related characteristics are used as the input of the layer, and a shadow reclassification result is obtained from random forests;
b.3, combining the non-shadow impervious surface obtained in the step S2.1 and the shadow impervious surface obtained in the step S2.2 to obtain impervious surface information in the whole area;
the classification of the water impermeable surface obtained in this example is shown in FIG. 4; wherein fig. 4 (a) is the result of the water-impermeable surface extraction algorithm proposed by the present invention, and fig. 4 (b) is the result of water-impermeable surface extraction obtained by the simple stacking algorithm; the method can effectively improve the underestimation of the impervious surface caused by the confusion of shadows and water bodies in urban impervious surface extraction; meanwhile, the distinguishing property of the bright impermeable surface and bare soil is improved at the edge of the urban area; thereby ensuring the extraction precision of the watertight surface in the city of large Fan Weicheng; the example adopts an Overall Accuracy (OA) and a Kappa coefficient to carry out quantitative Accuracy verification; by taking visual interpretation data of an image acquired in the latest time of Google Earth as a reference, and calculating the OA to be 93.9% and the Kappa coefficient to be 0.92, experiments prove that the invention can obtain an ideal impervious surface extraction result.
The foregoing is only a preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art, who is within the scope of the present invention, should make equivalent substitutions or modifications according to the technical scheme of the present invention and the inventive concept thereof, and should be covered by the scope of the present invention.
Claims (1)
1. The layered urban impervious surface extraction method based on the integration of optical and dual-polarized SAR is characterized by comprising the following steps of:
1) The image feature extraction based on the characteristic of the water-impermeable surface specifically comprises the following steps:
1.1, preprocessing optical and dual-polarized SAR images by adopting SNAP software respectively, wherein the preprocessing of the optical images comprises radiation correction and atmospheric correction; preprocessing the dual-polarized SAR image, including amplitude separation, stripping, radiation correction, speckle filtering and topography correction;
1.2, carrying out strict geographic registration on the preprocessed optical image and the dual-polarized SAR image by using EVNI software;
1.3, firstly calculating a normalized vegetation index NDVI and a normalized water body index for the pretreated optical image; then, respectively adopting gray level co-occurrence matrixes to calculate texture characteristics for visible light wave bands including a blue light wave band, a green light wave band, a red light wave band and a near infrared wave band;
1.4, obtaining polarization characteristics of the preprocessed dual-polarized SAR data through a covariance matrix, and respectively calculating texture characteristics of the polarized image by adopting a gray level co-occurrence matrix, wherein the method comprises the following steps of:
the polarization characteristics are selected from diagonal elements of covariance matrix C of the preprocessed dual-polarized SAR image, and C is defined as follows:
wherein S is VV Is a scattering matrix of VV polarization, S VH Is a VH polarized scattering matrix, x is defined as:
x＝[S VV S VH ] T
selecting a median, homogeneity and covariance of texture features, and respectively calculating each polarized image;
2) Urban impervious surface extraction based on layered framework, and the specific process comprises the following steps:
2.1, primarily classifying the images to obtain non-shadow vegetation, non-shadow impermeable water, non-shadow water body, non-shadow bare soil and shadow, stacking all optical and dual-polarization SAR features, and primarily classifying by using random forests;
2.2, selecting characteristics, and reclassifying the shadow obtained in the step S2.1 to obtain shadow vegetation, a shadow water body, shadow bare soil and a shadow impermeable water body: selecting dual-polarized SAR features and NDVI stacking input random forest, and reclassifying the shadows obtained in the step S2.1;
and 2.3, combining the non-shadow impermeable surface obtained in the step S2.1 and the shadow impermeable surface obtained in the step S2.2 to obtain impermeable surface information in the whole area.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111037677.9A CN113705523B (en) | 2021-09-06 | 2021-09-06 | Layered urban impervious surface extraction method based on optical and dual-polarized SAR fusion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111037677.9A CN113705523B (en) | 2021-09-06 | 2021-09-06 | Layered urban impervious surface extraction method based on optical and dual-polarized SAR fusion |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113705523A CN113705523A (en) | 2021-11-26 |
| CN113705523B true CN113705523B (en) | 2023-07-14 |
Family
ID=78660373
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111037677.9A Active CN113705523B (en) | 2021-09-06 | 2021-09-06 | Layered urban impervious surface extraction method based on optical and dual-polarized SAR fusion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113705523B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117095299B (en) * | 2023-10-18 | 2024-01-26 | 浙江省测绘科学技术研究院 | Grain crop extraction method, system, equipment and medium for crushing cultivation area |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103279951B (en) * | 2013-05-13 | 2016-03-09 | 武汉理工大学 | A kind of method of OO remote sensing image building and shadow extraction thereof |
| CN105930772A (en) * | 2016-04-13 | 2016-09-07 | 武汉大学 | City impervious surface extraction method based on fusion of SAR image and optical remote sensing image |
| CN108764326B (en) * | 2018-05-23 | 2021-06-22 | 北京工业大学 | City impervious stratum extraction method based on deep belief network |
| CN109919951B (en) * | 2019-03-14 | 2022-07-19 | 武汉大学 | Semantic-associated object-oriented urban impervious surface remote sensing extraction method and system |
-
2021
- 2021-09-06 CN CN202111037677.9A patent/CN113705523B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN113705523A (en) | 2021-11-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111242224B (en) | Multi-source remote sensing data classification method based on unmanned aerial vehicle extraction classification sample points | |
| CN105046087B (en) | A kind of Water-Body Information extraction method of remote sensing satellite multispectral image | |
| Sandmeier et al. | A physically-based model to correct atmospheric and illumination effects in optical satellite data of rugged terrain | |
| Laliberte et al. | Combining decision trees with hierarchical object-oriented image analysis for mapping arid rangelands | |
| US20210199579A1 (en) | Method and system for urban impervious surface extraction based on remote sensing | |
| CN101840581B (en) | Method for extracting profile of building from satellite remote sensing image | |
| CN108596103A (en) | High resolution ratio satellite remote-sensing image building extracting method based on optimal spectrum Index selection | |
| CN102622738B (en) | Method for recovering spectral information of hill shade area of Landsat thematic mapper/enhanced thematic mapper plus (TM/ETM+) image | |
| Blesius et al. | The use of the Minnaert correction for land‐cover classification in mountainous terrain | |
| CN107145891A (en) | A kind of Clean water withdraw method and system based on remote sensing image | |
| CN103942769A (en) | Satellite remote sensing image fusion method | |
| CN107564016B (en) | A kind of the Multi-Band Remote Sensing Images segmentation and labeling method of integrally object light spectrum information | |
| CN112052757B (en) | Method, device, equipment and storage medium for extracting fire trace information | |
| CN103839267A (en) | Building extracting method based on morphological building indexes | |
| CN111553922A (en) | Automatic cloud detection method for satellite remote sensing image | |
| CN113705523B (en) | Layered urban impervious surface extraction method based on optical and dual-polarized SAR fusion | |
| CN109918531A (en) | A kind of the seeking method, apparatus and computer readable storage medium of mother drug plants | |
| CN109598202A (en) | A kind of object-based satellite image multi objective built-up areas extraction method | |
| CN111611965B (en) | Method for extracting land surface water body based on Sentinel-2 image | |
| CN105512619B (en) | A kind of impervious surface information extracting method based on layering knowledge | |
| Mishra et al. | Bathymetric mapping using IKONOS multispectral data | |
| CN115456886A (en) | Aviation remote sensing image shadow removing method based on deep learning and illumination model | |
| Grigillo et al. | Classification based building detection from GeoEye-1 images | |
| CN114092418A (en) | Shadow detection method for unmanned aerial vehicle image of farmland crop | |
| CN110634077B (en) | Method for calculating vegetation index based on mixed vegetation index |
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 | ||
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Cheng Ji Inventor after: Zhi jun Jiao Inventor after: Wang Kai Inventor after: Chen Yong Inventor before: Sun Genyun Inventor before: Cheng Ji Inventor before: Zhi jun Jiao Inventor before: Zhang Aizhu Inventor before: Wang Kai Inventor before: Chen Yong |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |