CN111126234B - Automatic ship identification method based on multi-source satellite images - Google Patents
Automatic ship identification method based on multi-source satellite images Download PDFInfo
- Publication number
- CN111126234B CN111126234B CN201911311665.3A CN201911311665A CN111126234B CN 111126234 B CN111126234 B CN 111126234B CN 201911311665 A CN201911311665 A CN 201911311665A CN 111126234 B CN111126234 B CN 111126234B
- Authority
- CN
- China
- Prior art keywords
- image
- ship
- sentienl
- satellite
- index
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- 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
- Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
- Y02A10/40—Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Image Processing (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention provides a ship automatic identification method based on multi-source satellite images, which comprises the following steps: firstly, preprocessing a Sentienl-1 microwave satellite image acquired by a microwave satellite to obtain a Sentienl-1 ship index image, and constructing a Sentienl-1 ship identification model to obtain a ship-non-ship binary image; secondly, processing an annual improvement normalization difference water body index image acquired by a Landsat-8 optical satellite to obtain a water mask image; then processing the gradient image of the digital elevation model acquired by the digital elevation model satellite to obtain a gradient mask image; and finally, performing mask operation on the ship-non-ship binary image by using the water mask image and the gradient mask image to finish ship identification. The invention provides a Sentinel-1ship remote sensing identification index, combines the advantages of an optical satellite image, a microwave satellite image and a DEM image, improves the ship identification precision, and realizes the automatic ship identification of multi-source satellite image coupling.
Description
Technical Field
The invention relates to the technical field of remote sensing target identification, in particular to an automatic ship identification method based on multi-source satellite images.
Background
The ship remote sensing identification has very prominent practical significance in the fields of fishery supervision, emergency rescue, disaster relief and the like, and has extremely important scientific value in the field of satellite remote sensing application. For example, in the no-fish period or no-navigation area, the satellite remote sensing ship automatic identification technology is applied, and particularly on a large area scale, illegal ships can be effectively and quickly found. Compared with manual inspection, the working efficiency is greatly improved, and the cost of manpower and material resources is saved.
Due to the influence of the size of the ship, the requirement of remote sensing identification of the ship on the spatial resolution of the satellite image is high, and the image with the spatial resolution lower than 30 meters generally does not have the capacity of identifying the ship. In addition, ships are generally distributed in areas with more water areas, and the areas are rainy, so that optical image data is easy to lose, and the defects of the optical remote sensing method are exposed. The Sentinel-1 satellite image is a synthetic aperture radar microwave image, is not influenced by weather, and can work for 24 hours all weather. The space resolution of the Sentinel-1 satellite image is 10 meters, the time resolution is 6 days, the Sentinel-1 satellite image comprises two polarized images of VH and VV, and the Sentinel-1 satellite image has the potential of ship remote sensing identification.
Disclosure of Invention
Aiming at the defects in the background technology, the invention provides an automatic ship identification method based on multi-source satellite images, and solves the technical problem of poor identification capability caused by optical image data deletion in the existing optical remote sensing technology.
The technical scheme of the invention is realized as follows:
a ship automatic identification method based on multi-source satellite images comprises the following steps:
s1, collecting N Sentienl-1 microwave satellite images by using a Sentienl-1 microwave satellite, and respectively preprocessing the N Sentienl-1 microwave satellite images to obtain a Sentienl-1 ship index image;
s2, constructing a pixel histogram of a ship and a non-ship according to the Sentienl-1 ship index image obtained in the step S1, and obtaining a threshold value alpha for distinguishing the ship from the non-ship according to the pixel histogram;
s3, according to the N Sentienl-1 microwave satellite images, counting image characteristics of the ship in the Sentienl-1VH polarization images to obtain a parameter delta for optimizing ship identification, and completing construction of a Sentienl-1 ship identification model;
s4, preprocessing a real-time image acquired by a Sentinel-1 microwave satellite to obtain a real-time Sentinel-1ship index image;
s5, inputting the real-time Sentienl-1 ship index image and the real-time Sentienl-1VH polarization image in the step S4 into a Sentienl-1 ship identification model to obtain a ship-non-ship binary image;
s6, acquiring an annual improvement normalization difference water body index image by using a Landsat-8 optical satellite, and carrying out binarization processing on the annual improvement normalization difference water body index image to obtain a water mask image;
s7, acquiring a gradient image of the digital elevation model by using the digital elevation model satellite, and performing binarization processing on the gradient image of the digital elevation model to obtain a gradient mask image;
and S8, performing mask operation on the ship-non-ship binary image in the step S5 by using the water mask image in the step S6 and the gradient mask image in the step S7, and completing ship identification.
The method for preprocessing the Sentienl-1 microwave satellite image comprises the following steps:
s11, setting the pixel value of the VV polarization image in the Sentienl-1 microwave satellite image, of which the backscattering coefficient is larger than-1, as-1;
s12, using SSI i =(β vv,i -β vh,i )/β vv,i Calculating the Sentienl-1 microwave satellite image ship index to obtain a Sentienl-1 ship index image, wherein SSI i Represents the pixel value, beta, of the pixel i in the Sentienl-1 ship index image vv,i Represents the backscattering coefficient, beta, of the pixel i in the Sentiniel-1 VV polarized image vh,i The backscattering coefficient of the pixel i in the Sentinel-1VH polarized image is shown, i is 1,2, …, and n is the number of pixels.
The method for inputting the real-time Sentienl-1 ship index image and the real-time Sentienl-1VH polarization image into the Sentienl-1 ship identification model to obtain the ship-non-ship binary image comprises the following steps: and taking the picture elements of the real-time Sentienl-1 ship index image, of which the pixel value is greater than a threshold value alpha and the pixel value in the VH polarization image is less than a threshold value delta as a ship, setting the pixel values corresponding to the ship to be 1, and setting the other pixel values to be 0 to obtain a ship-non-ship binary image.
The method for obtaining the water mask image by carrying out binarization processing on the annual improved normalized difference water body index image comprises the following steps:
s61, processing the annual improved normalized difference water body index image by utilizing a maximum synthesis technology to obtain an improved normalized difference water body index maximization image;
s62, taking the pixels with the pixel value larger than 0.1 in the improved normalized difference water body index maximization image as the water body, setting the pixel values corresponding to the water body to be 1, and setting the other pixel values to be 0 to obtain the water mask image.
The method for obtaining the gradient mask image by carrying out binarization processing on the gradient image of the digital elevation model comprises the following steps of: and taking the pixels with the gradient smaller than 1 in the gradient image of the digital elevation model as possible areas of the ship, setting the pixel values corresponding to the possible areas of the ship to be 1, and setting the other pixel values to be 0 to obtain a gradient mask image.
The beneficial effect that this technical scheme can produce.
(1) The invention provides a remote sensing identification Index of a Ship based on a Sentinel-1 microwave satellite image, namely an SSI (Sentinel-1Ship Index), which fully develops the potential of a Sentinel-1 image for identifying the Ship;
(2) the invention adopts the microwave satellite image, which can effectively avoid the interference of cloud and rain weather on the remote sensing identification of the ship;
(3) the invention couples the advantages of multi-source satellite data such as optical satellite images, microwave satellite images, DEM images and the like, effectively improves the precision of ship identification, and realizes the automatic identification of ships coupled by multi-source satellite images.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a ship identification result of the present invention; the region in the figure is in the Poyang lake part region, and the time phase of the Sentinel-1 is 8 and 19 days in 2019.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without inventive effort based on the embodiments of the present invention, are within the scope of the present invention.
As shown in fig. 1, an embodiment of the present invention provides an automatic ship identification method based on multi-source satellite images, which includes the following specific steps:
s1, collecting N-11 Sentienl-1 microwave satellite images by using a Sentienl-1 microwave satellite, and respectively preprocessing the 11 Sentienl-1 microwave satellite images to obtain a Sentinell-1 Ship Index (SSI) image. The pretreatment method comprises the following steps:
s11, setting the pixel value of the VV polarization image in the Sentienl-1 microwave satellite image, of which the backscattering coefficient is larger than-1, as-1;
s12, analyzing the image characteristics of the ship through 1537 random sampling pixels based on the Sentienl-1 image to construct a Sentienl-1 ship index with the expression SSI i =(β vv,i -β vh,i )/β vv,i Wherein, SSI i Represents the pixel value, beta, of the pixel i in the Sentinel-1ship index image vv,i Represents the backscattering coefficient of the pixel i in the Sentiniel-1 VV polarized image, and the unit is dB, beta vh,i Representing the backscattering coefficient of the pixel i in the Sentinel-1VH polarization image, wherein the unit is dB, i is 1,2, …, n and n are the number of the pixels;
s13, calculating by using the SSI expression obtained in the step S12 to obtain a Sentienl-1 ship index image.
S2, constructing a pixel histogram of a ship and a non-ship according to the Sentinel-1ship index image obtained in the step S1, and obtaining a threshold value alpha for distinguishing the ship from the non-ship according to the pixel histogram; wherein the threshold α is-1.5. 1000 ship and non-ship pixel samples are respectively collected from the Sentinel-1ship index image obtained in the step S1, a ship-non-ship pixel histogram is constructed, and according to the distribution of pixels in the Sentinel-1ship index histogram, a boundary point between a ship pixel value and other pixel values can be obtained, wherein the boundary point is a threshold value alpha, and the threshold value alpha is one of parameters of a Sentinel-1ship identification model.
S3, counting image features of the ship in the Sentienl-1VH polarization image, and accordingly obtaining a parameter delta for optimizing ship identification; where δ is-19.5 dB. 2538 ship pixel samples are collected from the Sentienl-1VH polarization image obtained in step S1, the pixel value distribution range of the ship in the VH polarization image is counted, and the upper limit of the pixel value distribution range is taken as a second parameter of the Sentinel-1ship identification model, namely, a threshold value delta.
S4, preprocessing the real-time image acquired by the Sentinel-1 microwave satellite according to the processing method in the step S1 to obtain a real-time Sentinel-1ship index image.
S5, inputting all pixel values and VH polarization images of the real-time Sentinel-1ship index image in the step S4 into a Sentinel-1ship identification model to obtain a ship-non-ship binary image; and taking the pixels with pixel values larger than a threshold value alpha in the real-time Sentinel-1ship index image and pixel values smaller than a threshold value delta in the real-time VH polarization image as a ship, setting the pixel values corresponding to the ship to be 1, and setting the other pixel values to be 0 to obtain a ship-non-ship binary image.
S6, acquiring an annual improved normalized difference water body index image by using a Landsat-8 optical satellite, and processing the annual improved normalized difference water body index image to obtain a water mask image, wherein the water mask image is used for limiting a ship identification area in an area with water body distribution; the specific method comprises the following steps:
s61, processing the annual improved Normalized Difference Water Index (MNDWI) image by utilizing a maximum synthesis technology to obtain an improved Normalized Difference Water Index maximization image; the maximum value synthesis technology refers to comparing pixel values at the same position in a group of images, reserving the maximum pixel value, and sequentially comparing until pixel values at all positions are traversed.
S62, taking the pixels with the pixel value larger than 0.1 in the improved normalized difference water body index maximization image as the water body, setting the pixel values corresponding to the water body to be 1, setting the other pixel values to be 0, obtaining a water mask image, and further obtaining the annual maximum range of water body distribution.
S7, acquiring Digital Elevation Model (DEM) data by using a Digital Elevation model satellite, generating a gradient image by using the DEM data, and performing binarization processing on the gradient image to obtain a gradient mask image; the binarization processing method comprises the following steps: and taking the pixels with the gradient smaller than 1 in the gradient image as possible areas of the ship, setting the pixel values corresponding to the possible areas of the ship to be 1, and setting the other pixel values to be 0 to obtain the gradient mask image. Influenced by factors such as mountain shadow and the like, the water body mask image obtained in the step S6 has a phenomenon of error identification of the water body, and the deficiency of the water body mask can be effectively made up by using the gradient mask image.
And S8, performing mask operation on the ship-non-ship binary image in the step S5 by using the water mask image in the step S6 and the gradient mask image in the step S7 to eliminate the phenomenon that the onshore target is wrongly identified as a ship, wherein the pixel with the pixel value of 1 in the image after the mask operation is the ship, and the pixel with the pixel value of 0 is the non-ship, so that the ship identification is completed.
In order to verify the effect of the method, the Poyang lake area with 8-19 th month in 2019 is identified, the identification result is shown in fig. 2, the size of the plaque of the identification result is not the actual size of the ship and is the position of the ship (group) extracted by the method, and fig. 2 shows that the method can improve the precision of ship identification and realize automatic ship identification through multi-source satellite image coupling.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (4)
1. A ship automatic identification method based on multi-source satellite images is characterized by comprising the following steps:
s1, collecting N Sentienl-1 microwave satellite images by using a Sentienl-1 microwave satellite, and respectively preprocessing the N Sentienl-1 microwave satellite images to obtain a Sentienl-1 ship index image; the specific method comprises the following steps:
s11, setting the pixel value of the VV polarization image in the Sentienl-1 microwave satellite image, of which the backscattering coefficient is larger than-1, as-1;
s12, using SSI i =(β vv,i -β vh,i )/β vv,i Calculating the Sentienl-1 microwave satellite image ship index to obtain a Sentienl-1 ship index image, wherein SSI i Represents the pixel value, beta, of the pixel i in the Sentienl-1 ship index image vv,i Represents the backscattering coefficient, beta, of the pixel i in the Sentiniel-1 VV polarized image vh,i Representing the backscattering coefficient of a pixel i in a Sentinel-1VH polarization image, wherein i is 1,2, …, n and n are the number of the pixels;
s2, constructing a pixel histogram of a ship and a non-ship according to the Sentienl-1 ship index image obtained in the step S1, and obtaining a threshold value alpha for distinguishing the ship from the non-ship according to the pixel histogram;
s3, according to the N Sentienl-1 microwave satellite images, counting image characteristics of the ship in the Sentienl-1VH polarization images to obtain a parameter delta for optimizing ship identification, and completing construction of a Sentienl-1 ship identification model;
s4, preprocessing a real-time image acquired by a Sentiel-1 microwave satellite to obtain a real-time Sentienl-1 ship index image;
s5, inputting the real-time Sentienl-1 ship index image and the real-time Sentienl-1VH polarization image in the step S4 into a Sentienl-1 ship identification model to obtain a ship-non-ship binary image;
s6, acquiring an annual improvement normalization difference water body index image by using a Landsat-8 optical satellite, and carrying out binarization processing on the annual improvement normalization difference water body index image to obtain a water mask image;
s7, acquiring a gradient image of the digital elevation model by using the digital elevation model satellite, and performing binarization processing on the gradient image of the digital elevation model to obtain a gradient mask image;
and S8, performing mask operation on the ship-non-ship binary image in the step S5 by using the water mask image in the step S6 and the gradient mask image in the step S7, and completing ship identification.
2. The method for automatically identifying ships based on multisource satellite images according to claim 1, wherein the method for inputting the real-time Sentienl-1 ship index image and the real-time Sentienl-1VH polarization image into a Sentienl-1 ship identification model to obtain a ship-non-ship binary image comprises the following steps: and taking the picture elements of the real-time Sentienl-1 ship index image, of which the pixel value is greater than a threshold value alpha and the pixel value in the VH polarization image is less than a threshold value delta as a ship, setting the pixel values corresponding to the ship to be 1, and setting the other pixel values to be 0 to obtain a ship-non-ship binary image.
3. The automatic ship identification method based on the multisource satellite image according to claim 1, wherein the method for obtaining the water mask image by performing binarization processing on the annual improved normalized difference water body index image comprises the following steps:
s61, processing the annual improved normalized difference water body index image by utilizing a maximum synthesis technology to obtain an improved normalized difference water body index maximization image;
s62, taking the pixels with the pixel value larger than 0.1 in the improved normalized difference water body index maximization image as the water body, setting the pixel values corresponding to the water body to be 1, and setting the other pixel values to be 0 to obtain the water mask image.
4. The automatic ship identification method based on the multi-source satellite image according to claim 1, wherein the method for obtaining the gradient mask image by performing binarization processing on the gradient image of the digital elevation model comprises the following steps: pixels with the gradient smaller than 1 in the gradient image of the digital elevation model are used as areas where ships may exist, pixel values corresponding to the areas where the ships may exist are all set to be 1, and the rest pixel values are all set to be 0, so that a gradient mask image is obtained.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911311665.3A CN111126234B (en) | 2019-12-18 | 2019-12-18 | Automatic ship identification method based on multi-source satellite images |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911311665.3A CN111126234B (en) | 2019-12-18 | 2019-12-18 | Automatic ship identification method based on multi-source satellite images |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111126234A CN111126234A (en) | 2020-05-08 |
CN111126234B true CN111126234B (en) | 2022-08-09 |
Family
ID=70499772
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201911311665.3A Active CN111126234B (en) | 2019-12-18 | 2019-12-18 | Automatic ship identification method based on multi-source satellite images |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111126234B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111914624B (en) * | 2020-06-17 | 2023-08-22 | 交通运输部天津水运工程科学研究所 | Method for identifying ship illegal closing AIS (automatic identification system) behavior by utilizing high-resolution satellite image |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2610636A1 (en) * | 2011-12-29 | 2013-07-03 | Windward Ltd. | Providing near real-time maritime insight from satellite imagery and extrinsic data |
CN105046087A (en) * | 2015-08-04 | 2015-11-11 | 中国资源卫星应用中心 | Water body information automatic extraction method for multi-spectral image of remote sensing satellite |
-
2019
- 2019-12-18 CN CN201911311665.3A patent/CN111126234B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2610636A1 (en) * | 2011-12-29 | 2013-07-03 | Windward Ltd. | Providing near real-time maritime insight from satellite imagery and extrinsic data |
CN105046087A (en) * | 2015-08-04 | 2015-11-11 | 中国资源卫星应用中心 | Water body information automatic extraction method for multi-spectral image of remote sensing satellite |
Non-Patent Citations (3)
Title |
---|
A Novel Recognition Approach Based on Multi-agent;Yong-mei Zhang et al.;《IEEE Xplore》;20110124;全文 * |
Multi-targets recognition for surface moving platform vision system based on combined features;Zhongli Ma et al.;《IEEE Xplore》;20140828;全文 * |
基于多源遥感卫星的海面舰船目标检测方法;孙越娇等;《激光与红外》;20180220(第02期);全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN111126234A (en) | 2020-05-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108428220B (en) | Automatic geometric correction method for ocean island reef area of remote sensing image of geostationary orbit satellite sequence | |
CN110703244B (en) | Method and device for identifying urban water body based on remote sensing data | |
KR20200063682A (en) | Method and appartus for estimating stream flow discharge using satellite images at streams | |
CN110619647B (en) | Method for positioning fuzzy region of image based on combination of edge point frequency domain and spatial domain characteristics | |
CN113191374B (en) | PolSAR image ridge line extraction method based on pyramid attention network | |
CN108230375A (en) | Visible images and SAR image registration method based on structural similarity fast robust | |
CN110334623B (en) | Method for extracting collapsing information based on Sentinel-2A satellite remote sensing image | |
CN111104850A (en) | Remote sensing image building automatic extraction method and system based on residual error network | |
CN111126234B (en) | Automatic ship identification method based on multi-source satellite images | |
CN107133937B (en) | A kind of self-adapting enhancement method of infrared image | |
CN110849821A (en) | Black and odorous water body remote sensing identification method based on Bayesian theorem | |
CN108931825A (en) | A kind of remote sensing image clouds thickness detecting method based on atural object clarity | |
CN109542932B (en) | Customized screening method for Landsat-8 satellite selection remote sensing data set | |
CN112949657B (en) | Forest land distribution extraction method and device based on remote sensing image texture features | |
CN112697218B (en) | Reservoir capacity curve reconstruction method | |
CN117451012A (en) | Unmanned aerial vehicle aerial photography measurement method and system | |
CN113570554A (en) | Single image visibility detection method based on scene depth | |
CN107832805B (en) | Technology for eliminating influence of spatial position error on remote sensing soft classification precision evaluation based on probability position model | |
McCarthy et al. | Digital analysis of lichen cover: a technique for use in lichenometry and licnenology | |
CN115222837A (en) | True color cloud picture generation method and device, electronic equipment and storage medium | |
CN114580573A (en) | Image-based cloud amount, cloud shape and weather phenomenon inversion device and method | |
CN113435326A (en) | Reed spatial distribution identification method based on remote sensing and habitat characteristics | |
CN107967714A (en) | A kind of method that forest canopy density is automatically extracted by unmanned plane digital elevation model | |
CN112580504A (en) | Tree species classification counting method and device based on high-resolution satellite remote sensing image | |
CN103871065A (en) | Vegetation canopy layer aggregation effect quantitative evaluation method based on hemispherical videos |
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 |