CN113393497A - Ship target tracking method, device and equipment of sequence remote sensing image under condition of broken clouds - Google Patents

Ship target tracking method, device and equipment of sequence remote sensing image under condition of broken clouds Download PDF

Info

Publication number
CN113393497A
CN113393497A CN202110767749.9A CN202110767749A CN113393497A CN 113393497 A CN113393497 A CN 113393497A CN 202110767749 A CN202110767749 A CN 202110767749A CN 113393497 A CN113393497 A CN 113393497A
Authority
CN
China
Prior art keywords
ship target
image
target
data
remote sensing
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202110767749.9A
Other languages
Chinese (zh)
Other versions
CN113393497B (en
Inventor
姚力波
路兴强
林迅
王海鹏
张杨
孙炜玮
张翔宇
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Naval Aeronautical University
Original Assignee
Naval Aeronautical University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Naval Aeronautical University filed Critical Naval Aeronautical University
Priority to CN202110767749.9A priority Critical patent/CN113393497B/en
Publication of CN113393497A publication Critical patent/CN113393497A/en
Application granted granted Critical
Publication of CN113393497B publication Critical patent/CN113393497B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/20—Analysis of motion
    • G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/136—Segmentation; Edge detection involving thresholding
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
    • 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
    • G06T2207/10016—Video; Image sequence
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10032—Satellite or aerial image; Remote sensing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • 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
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30241—Trajectory
    • 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)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Multimedia (AREA)
  • Image Processing (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

The invention relates to a ship target tracking method, a ship target tracking device and ship target tracking equipment of a sequence remote sensing image under a clouding condition, belonging to the technical field of satellite data processing, wherein the method is characterized in that a remote sensing sequence image is obtained based on a preset satellite, and the remote sensing sequence image is preprocessed and detected to obtain a ship target; based on the ship target, performing combined positioning and correction on the ship target position according to the high-precision coastline data and the AIS data to obtain an object space coordinate of the ship target; and tracking the ship target according to the object space coordinates of the ship target, establishing a ship target model, performing track filtering by using a geographical coordinate transfer equation based on a midsplit navigation method and adopting a PDA frame, completing multi-frame association and track extraction, and acquiring the ship target track. According to the invention, the remote sensing image ship target can be directly tracked based on the measured data, the false alarm rate of the ship target under the condition of cloud breaking is reduced, and the accuracy of target tracking is improved.

Description

Ship target tracking method, device and equipment of sequence remote sensing image under condition of broken clouds
Technical Field
The invention belongs to the technical field of satellite data processing, and particularly relates to a method, a device and equipment for tracking a ship target by using a sequence remote sensing image under a clouding condition.
Background
The target tracking is to detect, extract, identify and track a moving target in a sequence image to obtain target motion parameters, so as to perform subsequent processing and analysis, thereby realizing behavior understanding of the moving target, and the perception of the target situation and the judgment of the overall situation can be directly influenced by the error of the target tracking.
Staring satellites are used as new generation observation satellites, and by means of the characteristic that the staring satellites can continuously observe a certain area, more dynamic information can be obtained compared with the traditional satellites. The high-orbit staring satellite adopts a geosynchronous orbit and can realize target monitoring in a large range. For example, the resolution of the high-resolution four-satellite is 50 meters, the ship target generally only occupies a few to a dozen pixels, the high-resolution four-satellite belongs to a weak target, the number of obtainable features is small, the optical satellite is susceptible to cloud layer interference, and meanwhile, the size of the clouded object is close to that of the ship target, so that high false alarm is easy to occur, and subsequent processing analysis is influenced.
Therefore, how to realize the tracking of the ship target of the sequence remote sensing image under the condition of the cloud breaking becomes a technical problem to be solved urgently in the prior art.
Disclosure of Invention
The invention provides a ship target tracking method, a ship target tracking device and ship target tracking equipment based on sequence remote sensing images under a clouding condition, so that the false alarm rate of ship targets under the clouding condition is reduced, and the accuracy of target tracking is improved.
The technical scheme provided by the invention is as follows:
on one hand, the ship target tracking method of the sequence remote sensing image under the condition of the cloud fragmentation comprises the following steps:
acquiring a remote sensing sequence image based on a preset satellite, and preprocessing and detecting the remote sensing sequence image to obtain a ship target;
based on the ship target, performing combined positioning and correction on the ship target position according to the high-precision coastline data and the AIS data to obtain an object space coordinate of the ship target;
and tracking the ship target according to the object space coordinates of the ship target, establishing a ship target model, performing track filtering by using a geographical coordinate transfer equation based on a midsplit navigation method and adopting a PDA frame, completing multi-frame association and track extraction, and acquiring the ship target track.
Optionally, the preprocessing and detecting the remote sensing sequence image to obtain a ship target includes:
preprocessing the remote sensing sequence image, performing sea-land segmentation and cloud region masking, and performing image registration by using a phase correlation method according to the characteristic that the transformation between each frame image of the sequence remote sensing image is approximate to translation transformation to obtain registration data;
according to the double-parameter constant false alarm rate detection, based on the set condition that sea clutter obeys Gaussian distribution, performing target coarse detection on the registration data in a sliding window mode to obtain coarse detection data;
and learning the characteristics of the ship target and the clouding target based on a convolutional neural network by using the constructed ship clouding target data set, detecting the coarse detection data one by one, eliminating the clouding target, obtaining the ship target and the centroid coordinate thereof, and acquiring fine detection data.
Optionally, the preprocessing the remote sensing sequence image, and performing sea-land segmentation and cloud region masking includes:
if the remote sensing sequence image is a single-waveband image, directly extracting land by adopting threshold segmentation;
and if the remote sensing sequence image is a multiband image, carrying out land distinguishing by utilizing spectral characteristics.
Optionally, the jointly locating and correcting the position of the ship target according to the high-precision coastline data and the AIS data based on the ship target includes:
and if the image of the ship target contains land, fusing the targets respectively obtained by correcting the high-precision coastline data and the AIS data, and if the image of the ship target does not contain land, correcting by using the AIS data as the dynamic GCP.
Optionally, the jointly positioning and correcting the position of the ship target according to the high-precision coastline data and the AIS data based on the ship target to obtain the object coordinates of the ship target includes:
when the image of the ship target contains land, acquiring an image-space map layer of a coastline based on RPC (remote procedure call) conversion, and acquiring a land binary image through a mask based on the image-space map layer; carrying out consistency purification on the image side map layer through a random sampling consistency algorithm, and calculating an error correction coefficient to obtain an object side coordinate corrected by a coastline;
screening AIS data based on the RPC file of the preset satellite to obtain screening data, calculating screening data image side coordinates according to the screening data, performing matching association on the screening data image side coordinates and a ship target detection result by using an iteration nearest point and a global nearest neighbor, screening partial association points through RANSAC to calculate RPC model image side error compensation parameters, and obtaining corrected object side coordinates;
and performing confidence evaluation on the object space coordinate corrected by the coastline and the object space coordinate corrected by the AIS to obtain the corrected confidence of the coastline and the AIS, and obtaining the finally corrected target object space coordinate by adopting a data fusion method based on the DS evidence theory.
Optionally, the tracking the ship target according to the object coordinates of the ship target, establishing a ship target model, performing track filtering by using a geographical coordinate transfer equation based on a midsplit navigation method and using a PDA frame, completing multi-frame association and track extraction, and obtaining the ship target track includes:
the PDA is used as a frame to carry out ship target modeling, and ship target tracking preposition work is completed;
and performing track management by adopting a logic method, establishing association constraint, completing multi-frame target association and tracking, and obtaining the ship target track.
In another aspect, a ship target tracking device for sequence remote sensing images under a cloud breaking condition comprises: the device comprises a processing detection module, a positioning correction module and a tracking module;
the processing and detecting module is used for acquiring a remote sensing sequence image based on a preset satellite, and preprocessing and detecting the remote sensing sequence image to obtain a ship target;
the positioning correction module is used for carrying out combined positioning and correction on the position of the ship target according to the high-precision coastline data and the AIS data based on the ship target to obtain an object coordinate of the ship target;
the tracking module is used for tracking the ship target according to the object space coordinates of the ship target, establishing a ship target model, using a geographical coordinate transfer equation based on a midsplit latitude navigation method, adopting a PDA frame to carry out track filtering, completing multiframe association and track extraction, and obtaining the ship target track.
Optionally, the processing and detecting module is configured to pre-process the remote sensing sequence image, perform sea-land segmentation and cloud region masking, perform image registration by using a phase correlation method according to a characteristic that a transformation between frame images of the sequence remote sensing image is approximate to a translation transformation, and acquire registration data; according to the double-parameter constant false alarm rate detection, based on the set condition that sea clutter obeys Gaussian distribution, performing target coarse detection on the registration data in a sliding window mode to obtain coarse detection data; and learning the characteristics of the ship target and the clouding target based on a convolutional neural network by using the constructed ship clouding target data set, detecting the coarse detection data one by one, eliminating the clouding target, obtaining the ship target and the centroid coordinate thereof, and acquiring fine detection data.
Optionally, the positioning and correcting module is configured to fuse the targets obtained by respectively correcting the high-precision coastline data and the AIS data when the image of the ship target contains land; when land is not present, corrections are made using AIS data as a dynamic GCP.
In another aspect, a ship target tracking device for sequence remote sensing images under a cloud breaking condition includes: a processor, and a memory coupled to the processor;
the memory is used for storing a computer program, and the computer program is at least used for executing the ship target tracking method of the sequence remote sensing images under the cloud breaking condition;
the processor is used for calling and executing the computer program in the memory.
The invention has the beneficial effects that:
according to the ship target tracking method, device and equipment of the sequence remote sensing image under the cloud breaking condition, the remote sensing sequence image is obtained based on the preset satellite, and the remote sensing sequence image is preprocessed and detected to obtain the ship target; based on the ship target, performing combined positioning and correction on the ship target position according to the high-precision coastline data and the AIS data to obtain an object space coordinate of the ship target; and tracking the ship target according to the object space coordinates of the ship target, establishing a ship target model, performing track filtering by using a geographical coordinate transfer equation based on a midsplit navigation method and adopting a PDA frame, completing multi-frame association and track extraction, and acquiring the ship target track. The ship target tracking method based on the sequence remote sensing image can directly track the ship target of the remote sensing image based on the measured data, relieves the overhigh false alarm rate of the scene target with more broken clouds, improves the tracking accuracy, has the advantages of accurate tracking, wide application range, good practical effect and the like, and can be directly used for relieving the problem of broken cloud scene target tracking in the actual satellite remote sensing data processing. A multi-dimensional and multi-stage ship target tracking algorithm is designed, so that the false alarm rate of the ship target under the condition of cloud breaking can be reduced, and the target tracking accuracy is improved.
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 schematic flow diagram of a ship target tracking method of a sequence remote sensing image under a clouding condition according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a dual-parameter CFAR sliding window according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of an image correction process according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of a ship target tracking device for sequence remote sensing images under a clouding condition according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of ship target tracking equipment for sequence remote sensing images under a clouding condition according to an embodiment of the present invention.
Detailed description of the invention
In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be described in detail below. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the examples given herein without any inventive step, are within the scope of the present invention.
In order to at least alleviate the technical problems provided by the invention, the embodiment of the invention provides a ship target tracking method of a sequence remote sensing image under a clouding condition.
Fig. 1 is a schematic flow diagram of a ship target tracking method using sequence remote sensing images under a cloud-breaking condition according to an embodiment of the present invention, and referring to fig. 1, the method according to the embodiment of the present invention may include the following steps:
and S1, acquiring a remote sensing sequence image based on a preset satellite, and preprocessing and detecting the remote sensing sequence image to obtain a ship target.
In the embodiment of the invention, the high-resolution four-number optical satellite can be defined as a preset satellite, and the remote sensing sequence image is obtained from the high-resolution four-number optical satellite, so that ship target tracking of the sequence remote sensing image under the condition of cloud breaking is carried out. It should be noted that in the embodiment of the present invention, only the optical satellite with the high resolution of four is taken as an example, which is not limited, and the ship target tracking of the sequence remote sensing image under the cloud breaking condition may also be performed based on other satellites, which is not described herein again.
For example, image preprocessing, target coarse detection and target fine detection of a cloud-breaking scene can be performed on each frame of image of the high-resolution four-number optical satellite sequence image respectively to obtain a ship target.
In some embodiments, optionally, preprocessing and detecting the remote sensing sequence image to obtain the ship target may include the following steps:
s11, preprocessing the remote sensing sequence images, carrying out sea-land segmentation and cloud region masking, and carrying out image registration by using a phase correlation method according to the characteristic that the transformation between the frame images of the high-resolution four-sequence remote sensing images is approximate to translation transformation to obtain registration data.
In some embodiments, optionally, the preprocessing the remote sensing sequence image, and performing sea-land segmentation and cloud region masking includes: if the remote sensing sequence image is a single-waveband image, directly extracting land by adopting threshold segmentation; and if the remote sensing sequence image is a multiband image, carrying out land distinguishing by utilizing the spectral characteristics.
For example, in this embodiment, image preprocessing may be performed on the top four optical satellite remote sensing sequence image, sea-land segmentation and cloud region masking are performed, and according to the characteristic that the transformation between each frame image of the top four optical satellite remote sensing sequence image is approximate to the translation transformation, the image registration may be performed by using a phase correlation method, which specifically includes the following steps:
and S111, sea and land segmentation and cloud area masking are carried out, and the sea and the land are segmented. The reflectivity of land is higher compared to the sea surface. If the input is a single-waveband image, land is directly extracted by adopting threshold segmentation. If the input is a multiband image, the spectral characteristics are used for land discrimination. Wherein, the sea area is detected through the water body index, so as to extract the Region of Interest (ROI). Extracting ocean regions by using green light and Near Infrared (NIR) wave band data in GF-4 multispectral data and adopting Normalized Difference Water Index (NDWI),
Figure BDA0003151377080000051
where ρ isG,ρNIRRespectively representing the reflectivity of a green light wave band and the NIR wave band.
S112, setting a threshold value TwaterObtaining a binary image B of the ocean area,
Figure BDA0003151377080000052
and performing regional operation processing on the binary image obtained by segmentation according to the difference of the density distribution of the ship and the land. And (3) counting the size proportion of the 0 value in each block in the area, judging the area to be land if the area is too large (a judgment reference can be set according to the requirement, the area is judged to be too large and land if the judgment reference is exceeded, and the area is not specifically limited), setting all 0 values in the area, otherwise setting all 1 values in the area, and finally setting the binary image to be B'.
S113, partitioning the image, wherein the partition pixel size is 512 multiplied by 512, and the image partition containing the land area is reserved. Setting reference image partition I by phase correlation method1And image partition I to be stabilized2The translation relationship between the two is as follows:
I1(x,y)＝I2(x+Δx,y+Δy)
where x, y represent image pixel coordinates and Δ x, Δ y represent I1And I2The amount of translation in the x and y directions. Fourier transform is carried out to obtain
F2(u,v)＝e-j2π(u+Δx,v+Δy)F1(u,v)
Where F (·) represents a frequency domain transform and (u, v) represents frequency domain coordinates. Performing correlation calculation on the two frames of images to obtain
Figure BDA0003151377080000061
Wherein, F*(. cndot.) denotes the conjugate of F (-), e-j2π(uΔx+vΔy)Is transformed into an impulse function, i.e.
F-1{e-j2π(uΔx+vΔy)}＝δ(x-Δx,y-Δy)
The translation amount delta x and delta y between two frame images are estimated through the formula, and automatic and rapid estimation between frames is achieved.
And S12, performing target coarse detection on the registration data in a sliding window mode according to the set condition that the double-parameter constant false alarm detection obeys Gaussian distribution based on the sea clutter, and acquiring coarse detection data.
The Constant False Alarm Rate (CFAR) technique is a technique in which a radar system determines whether a target signal exists by discriminating between a signal output from a receiver and noise while keeping a False Alarm probability Constant. Referring to fig. 2, fig. 2 is a schematic view of a sliding window structure of a dual-parameter CFAR according to an embodiment of the present invention. The method comprises the following specific steps:
s121, setting window sizes of a target window, a protection window and a clutter window, wherein the target window is generally 2 times of the side length of a pixel occupied by the minimum ship target and is set to be 1, the protection window is generally 2 times of the side length of the pixel occupied by the maximum ship target and is set to be 2, the clutter window needs to be large enough to ensure that the background is not influenced by other targets, non-background targets and other factors, a user can set the window according to requirements, and the window size is not specifically limited.
S122, carrying out CFAR detection on the image according to a target judgment criterion of the double-parameter CFAR, wherein the specific criterion is as follows:
μT＞μB+KcfarσBis judged as the target
μT＜μB+KcfarσBIs judged as background
Wherein, muT,μBMean, σ, of the target window and clutter window, respectivelyBVariance of clutter window, KcfarTo control the false alarm rate, andand converting the detection result into a binary image.
S123, 8 connected domain marking is carried out on the detected target area by using morphological processing, targets are respectively extracted, overlarge false targets are removed according to the size of target pixels, the centroid coordinate of each target is used as the coordinate of an image space, the centroid coordinate of each target is used as the center, the area of 10 pixels around the target is extracted, and the area is made into a sample slice to be used for the fine detection of the ship target under the subsequent cloud breaking condition.
S13, learning the characteristics of the ship target and the clouding target based on the convolutional neural network by utilizing the constructed high-resolution four-size ship clouding target data set, detecting the coarse detection data one by one, eliminating the clouding target, obtaining the ship target and the centroid coordinate thereof, and acquiring fine detection data.
Among them, Convolutional Neural Network (CNN) is a prior art. The method specifically comprises the following steps:
s131, taking target slices extracted after CFAR detection as samples, where a positive sample set is composed of fragmented cloud target slices, and a negative sample set is composed of ship target slices, and considering that the number of negative sample ship target slices is small compared to the positive sample fragmented cloud target slices, a Synthetic Minority Over-sampling Technique (SMOTE) is adopted to increase Minority samples, and a SMOTE algorithm flow is as follows:
1) for each sample X in the minority sample X, calculating the distance from the sample X to all samples in the minority sample set to obtain k neighbor of the sample X, and setting a sampling proportion according to the sample imbalance proportion to determine a sampling multiplying power N;
2) calculating xiK neighbor samples of (a) and stored in a set Xik；
3) From the set XikIn randomly selecting sample xij,j∈{1,2,…,k}；
4) Generating a random number lambda belonging to [0,1 ];
5) synthesis of sample xijAnd xiNew sample x in betweennewEach attribute value x ofnew,attrAnd new sample xnewLogging inSet S, the formula is as follows:
xnew,attr＝xi,attr+(xij,attr-xi,attr)×λ
wherein x isi,attrRepresents a sample xiProperty value of xij,attrRepresents a sample xijThe attribute value of (2);
6) repeating the steps 3) to 5) N times;
7) repeat step 2) -step 6) | X | times.
S132, performing feature learning by using CNN, using an AlexNet structure, and reserving an input shape to effectively identify the correlation of pixels of the image in the height direction and the width direction, and reducing the parameter size by repeatedly calculating the same convolution kernel and the input at different positions through a sliding window.
S133, inputting the image slice into a trained AlexNet model, performing remote sensing image target classification to obtain a ship target sample slice, and using the ship target sample slice as follow-up ship target tracking.
And S2, based on the ship target, performing combined positioning and correction on the ship target position according to the high-precision coastline data and the AIS data, and acquiring the object space coordinates of the ship target.
The automatic identification system ais (automatic identification system) is a ship navigation device. The AIS data is ship navigation data.
In some embodiments, optionally, based on the ship target, jointly locating and correcting the ship target position according to the high-precision coastline data and the AIS data, includes: and if the image of the ship target contains land, fusing the targets respectively obtained by correcting the high-precision coastline data and the AIS data, and if the image does not contain land, correcting by using the AIS data as the dynamic GCP.
For example, the high-precision coastline data and the AIS data can be used for carrying out joint positioning and correction on the ship target position, and if the image contains land, targets obtained by respectively correcting the high-precision coastline data and the AIS data are fused; if the land is not included, the AIS data is used as a dynamic GCP (group Control Point), namely the AIS data is used as a dynamic Ground Control Point to carry out correction. Fig. 3 shows a process of image correction, and fig. 3 is a schematic diagram of an image correction process according to an embodiment of the present invention.
The method specifically comprises the following steps:
s21, when the image of the ship target contains land, obtaining an image-side map layer of a coastline based on RPC (remote procedure calls) conversion, and obtaining a land binary image through a mask based on the image-side map layer; and (4) carrying out consistency purification on the image side image layer through a random sampling consistency algorithm, and calculating an error correction coefficient to obtain an object side coordinate under coastline correction, namely a ship target position.
For example, land areas are extracted through image preprocessing, the rough position of the image area is obtained by utilizing a rational polynomial coefficient file, and a high-precision coastline image layer is screened from the rough position. The method comprises the following steps of obtaining an image-space map layer of a coastline by RPC conversion, obtaining a land binary image by a mask, carrying out consistency purification image space by a Random Sample Consensus (RANSAC) algorithm, and calculating an error correction coefficient so as to obtain a ship target position under coastline correction, wherein the method comprises the following specific steps:
s211, establishing an RPC model, wherein the definition of the RPC model is a ratio polynomial about ground point coordinates and image coordinates, and the method specifically comprises the following steps:
Figure BDA0003151377080000081
wherein, the form of Num and Den polynomial is:
p(u,v,w)＝a1+a2v+a3u+a4w+a5vu+a6vw+a7uw+a8v2+a9u2+a10w2+
a11uvw+a12v3+a13vu2+a14vw2+a15v2u+a16u3+a17uw2+a18v2w+a19u2w+a20w3
wherein, a1～a20For rational polynomial coefficients, the 4 polynomials in the RPC model have 80 coefficients, u, v, w, l, s are regularization coefficients, and
Figure BDA0003151377080000082
the coordinates of the ground, i.e. object space, respectively represent latitude, longitude and elevation, and (L, S) are the coordinates of the remote sensing image, i.e. image space, respectively are row numbers and column numbers, the regularization can be expressed as
Figure BDA0003151377080000083
l＝(L-l0)/ls,s＝(S-s0)/ss
In the formula (I), the compound is shown in the specification,
Figure BDA0003151377080000084
λ0,h0,l0,s0and
Figure BDA0003151377080000085
λs,hs,ls,ssrespectively, normalized compensation and scale parameters.
S212, introducing an affine model in an image space error compensation model, correcting system errors, performing position correction by using Global High-precision coastline data (GSHHS), performing RPC projection on the GSHHS within an image display range and masking a closed coastline to obtain a land area template image, performing matching by using Normalized autocorrelation (NCC), calculating NCC values of the template image and the land binary image in a certain search window, finding the position with the largest NCC, and forming a matching point set;
s213, further eliminating low-precision matching points by adopting a RANSAC algorithm, randomly selecting 5 pairs of matching points each time, calculating affine transformation parameters, then solving the error of other point pairs under current transformation, and taking the point pairs with the error smaller than a threshold value as interior points to obtain an interior point set.
Repeating the random sampling for N times to obtain a maximum inner point set, solving affine transformation parameters by using a Least square method (LS), and obtaining object space coordinates of the target after the coastline is corrected by using an RPC model
Figure BDA0003151377080000086
S22, screening AIS data based on the RPC files of the preset satellites to obtain screening data, calculating image coordinates of the screening data according to the screening data, performing matching association on the image coordinates of the screening data and a ship target detection result by using an iteration nearest point and a global nearest neighbor, screening partial association points through RANSAC to calculate RPC model image error compensation parameters, and obtaining corrected object coordinates.
For example, according to a related RPC file with a high score of four, AIS data is screened, image side coordinates of the AIS data are obtained by using the screened data, matching and associating the AIS image side data with a ship target detection result by using Iterative Closest Point (ICP) and Global Nearest Neighbor (GNN), and partial associated Point pairs are screened by RANSAC to calculate RPC model image side error compensation parameters, so as to obtain corrected object side coordinates, which specifically includes the following steps:
s221, by utilizing the imaging time and the approximate imaging range of the high-resolution four-number satellite, AIS data of similar time and space are screened out, and the same target data is stored in a time sequence through ship identification numbers. Performing linear interpolation and reasonable extrapolation on AIS data of the same target to obtain a target position corresponding to imaging time;
s222, according to the position information of the target, performing data coarse association by adopting iterative weighted Least Squares (IRLS) ICP (iterative weighted Least Squares), and adopting a Munkres distribution algorithm. Performing RPC correction by using the well-associated AIS data, and solving the object space coordinates of the target after AIS correction by using the image space affine transformation compensation model
Figure BDA0003151377080000091
And S23, performing confidence evaluation on the object space coordinates after the coastline correction and the object space coordinates after the AIS correction to obtain the confidence degrees of the correction of the coastline and the object space coordinates after the AIS correction, and obtaining the final corrected target object space coordinates by adopting a data fusion method based on a DS evidence theory.
For example, confidence evaluation is carried out on the object coordinate corrected by GSHHS and the object coordinate corrected by AIS to obtain the confidence of correction of the two, and the data fusion method based on DS evidence theory is adopted to obtain the final corrected target object coordinate
Figure BDA0003151377080000092
S3, tracking the ship target according to the object space coordinates of the ship target, establishing a ship target model, using a geographical coordinate transfer equation based on a midsplit navigation method, adopting a PDA frame to carry out track filtering, completing multiframe association and track extraction, and obtaining a ship target track.
In some embodiments, optionally, the following steps may be specifically included:
and S31, carrying out ship target modeling by taking the PDA as a frame, and carrying out ship target tracking preposition work.
For example, the method specifically comprises the following steps:
s311, using λi(k),
Figure BDA0003151377080000093
Respectively representing the position of the ith measurement at the time of target k, namely longitude and latitude, selecting a mid-latitude navigation method, and dividing the mid-latitude
Figure BDA0003151377080000094
Is approximated to
Figure BDA0003151377080000095
From a starting position
Figure BDA0003151377080000096
To the end position
Figure BDA0003151377080000097
The relative transformation relation of the target geographic coordinates is as follows:
Figure BDA0003151377080000098
wherein T is a time interval, l is the number of miles corresponding to longitude 1 degree arc length on the equator,
Figure BDA0003151377080000101
for the velocity components in the longitude and latitude directions, the state at the target time k is obtained as follows:
Figure BDA0003151377080000102
s312, the ith measurement at target time k is represented as:
Figure BDA0003151377080000103
by Z (k) representing the set of measurements within the target wave gate at time k, i.e.
Figure BDA0003151377080000104
mkFor the measurement in the target wave gate, set ZkIs a cumulative set of measurements taken before (including) time k, there
Figure BDA0003151377080000105
It is possible to obtain,
Zk＝Zk-1+ Z (k) formula (16).
S313, set thetai(k) To measure z at time ki(k) For the probability of the target measurement, i is 1,2, …, mkIn particular, when i is 0, θ0(k) The probability that no measurement is taken as the target measurement at time k is represented by ZkAs a condition, defining an interconnection probability betai(k)：
βi(k)＝Pr{θi(k)|Zk}
Using Bayes total probability formula to obtain
Figure BDA0003151377080000106
S314, calculating interconnection probability zi(k) When measured as a target measurement, the likelihood function is:
Figure BDA0003151377080000107
in the formula, PGIs the gate probability when zi(k) When not used as the target measurement, there are:
Figure BDA0003151377080000108
in the formula, VkIs the volume of the associated wave gate.
Probability P of being completely detected by correct measurementDProbability mass function versus event theta with spurious measurementsiSolving to obtain
Figure BDA0003151377080000109
Figure BDA0003151377080000111
Here, the definition
Figure BDA0003151377080000112
The state and covariance updates were performed using first order EKFs.
And S32, performing track management by adopting a logic method, establishing association constraint, completing multi-frame target association and tracking, and obtaining the ship target track.
The specific steps can be as follows:
s321, under the PDA frame, using χ with degree of freedom of 22Distributed random variables
Figure BDA0003151377080000113
Designing an elliptic wave gate if the measurement Z (k +1) of the target in the geographic coordinate system meets the requirement
Figure BDA0003151377080000114
The measurement is deemed to be within the target wave gate.
And S322, performing track management, wherein when the track is started, if 3 effective measurements exist in 4 scans, the temporary track is changed into a determined track, in the existing track, if no measurement is updated for L continuous moments from a certain moment, the track is considered to be terminated, if the number of the continuous moments is less than L, the track is considered to be missed for detection, and the state is updated by using a predicted value.
The ship target tracking method of the sequence remote sensing image under the condition of the cloud fragmentation provided by the embodiment of the invention comprises the following steps: acquiring a remote sensing sequence image based on a preset satellite, and preprocessing and detecting the remote sensing sequence image to obtain a ship target; based on the ship target, performing combined positioning and correction on the ship target position according to the high-precision coastline data and the AIS data to obtain an object coordinate of the ship target; according to the object space coordinates of the ship target, tracking the ship target, establishing a ship target model, using a geographical coordinate transfer equation based on a midsplit navigation method, adopting a PDA frame to carry out track filtering, completing multi-frame association and track extraction, and obtaining a ship target track. The ship target tracking method based on the sequence remote sensing images can directly track the ship target of the remote sensing images based on the measured data, the technical problem that the false alarm rate of scene targets with more broken clouds is too high is solved, the tracking accuracy is improved, the ship target tracking method based on the sequence remote sensing images has the advantages of being accurate in tracking, wide in application range, good in practical effect and the like, and the ship target tracking method based on the sequence remote sensing images can be directly used for solving the problem of broken cloud scene target tracking in actual satellite remote sensing data processing. A multi-dimensional and multi-stage ship target tracking algorithm is designed, so that the false alarm rate of the ship target under the condition of cloud breaking can be reduced, and the target tracking accuracy is improved.
Based on a general inventive concept, the embodiment of the invention also provides a ship target tracking device of the sequence remote sensing image under the condition of cloud breaking.
Fig. 4 is a schematic structural diagram of a ship target tracking apparatus for sequence remote sensing images under a cloud-breaking condition according to an embodiment of the present invention, and referring to fig. 4, the apparatus according to the embodiment of the present invention may include the following structures: a process detection module 41, a position correction module 42 and a tracking module 43.
The processing and detecting module 41 is configured to obtain a remote sensing sequence image based on a preset satellite, and perform preprocessing and detection on the remote sensing sequence image to obtain a ship target;
the positioning correction module 42 is used for carrying out combined positioning and correction on the position of the ship target according to the high-precision coastline data and the AIS data based on the ship target to obtain an object coordinate of the ship target;
and the tracking module 43 is configured to track a ship target according to the object coordinates of the ship target, establish a ship target model, perform track filtering by using a geographical coordinate transfer equation based on a midsplit navigation method and using a PDA frame, complete multiframe association and track extraction, and acquire a ship target track.
Optionally, the processing and detecting module 41 is configured to perform preprocessing on the remote sensing sequence image, perform sea-land segmentation and cloud region masking, perform image registration by using a phase correlation method according to a characteristic that a transformation between frame images of the sequence remote sensing image is approximate to a translation transformation, and acquire registration data; according to the double-parameter constant false alarm rate detection, based on the set condition that sea clutter obeys Gaussian distribution, performing target coarse detection on the registration data in a sliding window mode to obtain coarse detection data; and learning the characteristics of the ship target and the clouding target based on a convolutional neural network by using the constructed ship clouding target data set, detecting the coarse detection data one by one, removing the clouding target, obtaining the ship target and the centroid coordinate thereof, and acquiring fine detection data.
Optionally, the processing and detecting module 41 is configured to directly extract a land by using threshold segmentation when the remote sensing sequence image is a single-band image; when the remote sensing sequence image is a multiband image, the land distinguishing is carried out by utilizing the spectral characteristics.
Optionally, the positioning and correcting module is configured to fuse the targets obtained by respectively correcting the high-precision coastline data and the AIS data when the image of the ship target includes land; when land is not present, corrections are made using AIS data as a dynamic GCP.
Optionally, the positioning correction module 42 is configured to, when the image of the ship target contains land, obtain an image-space layer of the coastline based on RPC transformation, and obtain a land binary image through a mask based on the image-space layer; carrying out consistency purification on the image side map layer through a random sampling consistency algorithm, and calculating an error correction coefficient to obtain an object side coordinate corrected by a coastline; screening AIS data based on RPC files of preset satellites to obtain screening data, calculating image coordinates of the screening data according to the screening data, performing matching association between the image coordinates of the screening data and a ship target detection result by using an iteration nearest point and a global nearest neighbor, screening partial association points through RANSAC to calculate RPC model image error compensation parameters, and obtaining corrected object coordinates; and performing confidence evaluation on the object space coordinate corrected by the coastline and the object space coordinate corrected by the AIS to obtain the corrected confidence of the coastline and the AIS, and obtaining the finally corrected target object space coordinate by adopting a data fusion method based on the DS evidence theory.
Optionally, the tracking module 43 is configured to perform ship target modeling by using the PDA as a frame, and perform ship target tracking preposition work; and performing track management by adopting a logic method, establishing association constraint, completing multi-frame target association and tracking, and obtaining the ship target track.
With regard to the apparatus in the above-described embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.
The ship target tracking device for the sequence remote sensing image under the cloud breaking condition, provided by the embodiment of the invention, is used for acquiring the remote sensing sequence image based on the preset satellite, and preprocessing and detecting the remote sensing sequence image to obtain a ship target; based on the ship target, performing combined positioning and correction on the ship target position according to the high-precision coastline data and the AIS data to obtain an object coordinate of the ship target; according to the object space coordinates of the ship target, tracking the ship target, establishing a ship target model, using a geographical coordinate transfer equation based on a midsplit navigation method, adopting a PDA frame to carry out track filtering, completing multi-frame association and track extraction, and obtaining a ship target track. The ship target tracking method based on the sequence remote sensing images can directly track the ship target of the remote sensing images based on the measured data, the technical problem that the false alarm rate of scene targets with more broken clouds is too high is solved, the tracking accuracy is improved, the ship target tracking method based on the sequence remote sensing images has the advantages of being accurate in tracking, wide in application range, good in practical effect and the like, and the ship target tracking method based on the sequence remote sensing images can be directly used for solving the problem of broken cloud scene target tracking in actual satellite remote sensing data processing.
Based on a general inventive concept, the embodiment of the invention also provides ship target tracking equipment of the sequence remote sensing image under the condition of cloud breaking.
Fig. 5 is a schematic structural diagram of a ship target tracking device for sequence remote sensing images under a cloud-breaking condition according to an embodiment of the present invention, referring to fig. 5, the device according to the embodiment of the present invention includes: a processor 51, and a memory 52 connected to the processor.
The memory 52 is configured to store a computer program, where the computer program is at least used in the ship target tracking method based on the sequence remote sensing image under the cloud breaking condition described in any of the above embodiments;
the processor 51 is used to invoke and execute computer programs in the memory.
Embodiments of the present invention also provide a storage medium based on one general inventive concept.
A storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the ship target tracking method of the sequence remote sensing image under the cloud breaking condition are realized.
The above description is only for the specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present invention, and all the changes or substitutions should be covered within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the appended claims.
It is understood that the same or similar parts in the above embodiments may be mutually referred to, and the same or similar parts in other embodiments may be referred to for the content which is not described in detail in some embodiments.
It should be noted that the terms "first," "second," and the like in the description of the present invention are used for descriptive purposes only and are not to be construed as indicating or implying relative importance. Further, in the description of the present invention, the meaning of "a plurality" means at least two unless otherwise specified.
Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing specific logical functions or steps of the process, and alternate implementations are included within the scope of the preferred embodiment of the present invention in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present invention.
It should be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the various steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
It will be understood by those skilled in the art that all or part of the steps carried by the method for implementing the above embodiments may be implemented by hardware related to instructions of a program, which may be stored in a computer readable storage medium, and when the program is executed, the program includes one or a combination of the steps of the method embodiments.
In addition, functional units in the embodiments of the present invention may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a stand-alone product, may also be stored in a computer readable storage medium.
The storage medium mentioned above may be a read-only memory, a magnetic or optical disk, etc.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.

Claims (10)

1. The ship target tracking method of the sequence remote sensing image under the condition of cloud fragmentation is characterized by comprising the following steps of:
acquiring a remote sensing sequence image based on a preset satellite, and preprocessing and detecting the remote sensing sequence image to obtain a ship target;
based on the ship target, performing combined positioning and correction on the ship target position according to the high-precision coastline data and the AIS data to obtain an object space coordinate of the ship target;
and tracking the ship target according to the object space coordinates of the ship target, establishing a ship target model, performing track filtering by using a geographical coordinate transfer equation based on a midsplit navigation method and adopting a PDA frame, completing multi-frame association and track extraction, and acquiring the ship target track.
2. The method of claim 1, wherein the preprocessing and detecting the remote sensing sequence images to obtain a ship target comprises:
preprocessing the remote sensing sequence image, performing sea-land segmentation and cloud region masking, and performing image registration by using a phase correlation method according to the characteristic that the transformation between each frame image of the sequence remote sensing image is approximate to translation transformation to obtain registration data;
according to the double-parameter constant false alarm rate detection, based on the set condition that sea clutter obeys Gaussian distribution, performing target coarse detection on the registration data in a sliding window mode to obtain coarse detection data;
and learning the characteristics of the ship target and the clouding target based on a convolutional neural network by using the constructed ship clouding target data set, detecting the coarse detection data one by one, eliminating the clouding target, obtaining the ship target and the centroid coordinate thereof, and acquiring fine detection data.
3. The method of claim 2, wherein the preprocessing the remote sensing sequence image for sea-land segmentation and cloud region masking comprises:
if the remote sensing sequence image is a single-waveband image, directly extracting land by adopting threshold segmentation;
and if the remote sensing sequence image is a multiband image, carrying out land distinguishing by utilizing spectral characteristics.
4. The method of claim 1, wherein jointly locating and correcting a ship target location from high accuracy shoreline data and AIS data based on the ship target comprises:
and if the image of the ship target contains land, fusing the targets respectively obtained by correcting the high-precision coastline data and the AIS data, and if the image of the ship target does not contain land, correcting by using the AIS data as the dynamic GCP.
5. The method of claim 4, wherein the jointly locating and correcting the ship target position based on the ship target according to the high-precision shoreline data and the AIS data to obtain the object coordinates of the ship target comprises:
when the image of the ship target contains land, acquiring an image-space map layer of a coastline based on RPC (remote procedure call) conversion, and acquiring a land binary image through a mask based on the image-space map layer; carrying out consistency purification on the image side map layer through a random sampling consistency algorithm, and calculating an error correction coefficient to obtain an object side coordinate corrected by a coastline;
screening AIS data based on the RPC file of the preset satellite to obtain screening data, calculating screening data image side coordinates according to the screening data, performing matching association on the screening data image side coordinates and a ship target detection result by using an iteration nearest point and a global nearest neighbor, screening partial association points through RANSAC to calculate RPC model image side error compensation parameters, and obtaining corrected object side coordinates;
and performing confidence evaluation on the object space coordinate corrected by the coastline and the object space coordinate corrected by the AIS to obtain the corrected confidence of the coastline and the AIS, and obtaining the finally corrected target object space coordinate by adopting a data fusion method based on the DS evidence theory.
6. The method according to claim 1, wherein the tracking of the ship target according to the objective coordinates of the ship target, establishing a ship target model, performing track filtering using a geographical coordinate transfer equation based on a mid-latitude navigation method and using a PDA frame to complete multi-frame association and track extraction, and obtaining the ship target track comprises:
the PDA is used as a frame to carry out ship target modeling, and ship target tracking preposition work is completed;
and performing track management by adopting a logic method, establishing association constraint, completing multi-frame target association and tracking, and obtaining the ship target track.
7. Naval vessel target tracking means of sequence remote sensing image under the garrulous cloud condition, its characterized in that includes: the device comprises a processing detection module, a positioning correction module and a tracking module;
the processing and detecting module is used for acquiring a remote sensing sequence image based on a preset satellite, and preprocessing and detecting the remote sensing sequence image to obtain a ship target;
the positioning correction module is used for carrying out combined positioning and correction on the position of the ship target according to the high-precision coastline data and the AIS data based on the ship target to obtain an object coordinate of the ship target;
the tracking module is used for tracking the ship target according to the object space coordinates of the ship target, establishing a ship target model, using a geographical coordinate transfer equation based on a midsplit latitude navigation method, adopting a PDA frame to carry out track filtering, completing multiframe association and track extraction, and obtaining the ship target track.
8. The device according to claim 7, wherein the processing detection module is configured to pre-process the remote sensing sequence image, perform sea-land segmentation and cloud region masking, perform image registration by using a phase correlation method according to a characteristic that a transformation between frame images of the sequence remote sensing image approximates to a translation transformation, and acquire registration data; according to the double-parameter constant false alarm rate detection, based on the set condition that sea clutter obeys Gaussian distribution, performing target coarse detection on the registration data in a sliding window mode to obtain coarse detection data; and learning the characteristics of the ship target and the clouding target based on a convolutional neural network by using the constructed ship clouding target data set, detecting the coarse detection data one by one, eliminating the clouding target, obtaining the ship target and the centroid coordinate thereof, and acquiring fine detection data.
9. The apparatus of claim 7, wherein the positioning correction module is configured to fuse objects obtained by respectively correcting high-precision coastline data and AIS data when the image of the ship object contains land; when land is not present, corrections are made using AIS data as a dynamic GCP.
10. Naval vessel target tracking equipment of sequence remote sensing image under the garrulous cloud condition, its characterized in that includes: a processor, and a memory coupled to the processor;
the memory is used for storing a computer program, and the computer program is at least used for executing the ship target tracking method of the sequence remote sensing image under the clouding condition as claimed in any one of claims 1-6;
the processor is used for calling and executing the computer program in the memory.
CN202110767749.9A 2021-07-07 2021-07-07 Ship target tracking method, device and equipment of sequence remote sensing image under condition of broken clouds Active CN113393497B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110767749.9A CN113393497B (en) 2021-07-07 2021-07-07 Ship target tracking method, device and equipment of sequence remote sensing image under condition of broken clouds

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110767749.9A CN113393497B (en) 2021-07-07 2021-07-07 Ship target tracking method, device and equipment of sequence remote sensing image under condition of broken clouds

Publications (2)

Publication Number Publication Date
CN113393497A true CN113393497A (en) 2021-09-14
CN113393497B CN113393497B (en) 2022-08-09

Family

ID=77625435

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110767749.9A Active CN113393497B (en) 2021-07-07 2021-07-07 Ship target tracking method, device and equipment of sequence remote sensing image under condition of broken clouds

Country Status (1)

Country Link
CN (1) CN113393497B (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114898213A (en) * 2022-05-19 2022-08-12 中国人民解放军国防科技大学 AIS knowledge assistance-based remote sensing image rotating ship target detection method
CN114937216A (en) * 2022-07-25 2022-08-23 中国科学院空天信息创新研究院 Method, device, equipment and medium for stabilizing sequence images based on ship tracks
CN116188519A (en) * 2023-02-07 2023-05-30 中国人民解放军海军航空大学 Ship target motion state estimation method and system based on video satellite
CN116416136A (en) * 2023-04-17 2023-07-11 北京卫星信息工程研究所 Data amplification method for ship target detection of visible light remote sensing image and electronic equipment

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108898065A (en) * 2018-05-31 2018-11-27 北京航空航天大学 Candidate regions quickly screen and the depth network Ship Target Detection method of dimension self-adaption
CN111127506A (en) * 2019-11-19 2020-05-08 北京空间飞行器总体设计部 Sequence image-based marine moving target comprehensive detection method
CN111402299A (en) * 2020-04-08 2020-07-10 中国人民解放军海军航空大学 Remote sensing image target tracking method and device based on stationary orbit staring satellite
CN112686095A (en) * 2020-12-04 2021-04-20 中国人民解放军海军航空大学 Ship target track correlation method for stationary orbit staring satellite remote sensing image
CN112966635A (en) * 2021-03-19 2021-06-15 北京理工大学 Low-resolution time sequence remote sensing image-oriented moving ship detection method and device

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108898065A (en) * 2018-05-31 2018-11-27 北京航空航天大学 Candidate regions quickly screen and the depth network Ship Target Detection method of dimension self-adaption
CN111127506A (en) * 2019-11-19 2020-05-08 北京空间飞行器总体设计部 Sequence image-based marine moving target comprehensive detection method
CN111402299A (en) * 2020-04-08 2020-07-10 中国人民解放军海军航空大学 Remote sensing image target tracking method and device based on stationary orbit staring satellite
CN112686095A (en) * 2020-12-04 2021-04-20 中国人民解放军海军航空大学 Ship target track correlation method for stationary orbit staring satellite remote sensing image
CN112966635A (en) * 2021-03-19 2021-06-15 北京理工大学 Low-resolution time sequence remote sensing image-oriented moving ship detection method and device

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
刘勇: "多源卫星舰船目标观测信息融合技术研究", 《中国优秀博硕士学位论文全文数据库(博士)工程科技Ⅱ辑》 *
刘钰群 等: "光学遥感层次化智能解译船只目标检测技术", 《 第十四届全国信号和智能信息处理与应用学术会议论文集》 *
姚力波 等: "基于高分四号卫星的舰船目标跟踪", 《第四届高分辨率对地观测学术年会论文集》 *

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114898213A (en) * 2022-05-19 2022-08-12 中国人民解放军国防科技大学 AIS knowledge assistance-based remote sensing image rotating ship target detection method
CN114937216A (en) * 2022-07-25 2022-08-23 中国科学院空天信息创新研究院 Method, device, equipment and medium for stabilizing sequence images based on ship tracks
CN116188519A (en) * 2023-02-07 2023-05-30 中国人民解放军海军航空大学 Ship target motion state estimation method and system based on video satellite
CN116188519B (en) * 2023-02-07 2023-10-03 中国人民解放军海军航空大学 Ship target motion state estimation method and system based on video satellite
CN116416136A (en) * 2023-04-17 2023-07-11 北京卫星信息工程研究所 Data amplification method for ship target detection of visible light remote sensing image and electronic equipment
CN116416136B (en) * 2023-04-17 2023-12-19 北京卫星信息工程研究所 Data amplification method for ship target detection of visible light remote sensing image and electronic equipment

Also Published As

Publication number Publication date
CN113393497B (en) 2022-08-09

Similar Documents

Publication Publication Date Title
CN113393497B (en) Ship target tracking method, device and equipment of sequence remote sensing image under condition of broken clouds
US10521694B2 (en) 3D building extraction apparatus, method and system
Mason et al. Accurate and efficient determination of the shoreline in ERS-1 SAR images
Hebert Outdoor scene analysis using range data
Liu et al. Algorithmic foundation and software tools for extracting shoreline features from remote sensing imagery and LiDAR data
CN107025654B (en) SAR image self-adaptive ship detection method based on global iterative inspection
CN102252681A (en) Global positioning system (GPS) and machine vision-based integrated navigation and positioning system and method
Hahmann et al. Surface water body detection in high-resolution TerraSAR-X data using active contour models
US20210132214A1 (en) Synthetic aperture radar image analysis system, synthetic aperture radar image analysis method, and synthetic aperture radar image analysis program
JP2014523572A (en) Generating map data
CN112270675B (en) Urban waterlogging area detection method based on polarized radar remote sensing image
CN113743385A (en) Unmanned ship water surface target detection method and device and unmanned ship
CN115540908A (en) InSAR interference fringe matching method based on wavelet transformation
CN111145198A (en) Non-cooperative target motion estimation method based on rapid corner detection
Petković et al. An overview on horizon detection methods in maritime video surveillance
Pan et al. Macroalgae monitoring from satellite optical images using Context-sensitive level set (CSLS) model
CN117523568A (en) Bridge height detection method, device, equipment and medium based on unmanned ship
Barreto et al. Estimating accurate water levels for rivers and reservoirs by using SAR products: A multitemporal analysis
Zhang et al. A novel saliency-driven oil tank detection method for synthetic aperture radar images
Wang et al. Lidar ground filtering algorithm for urban areas using scan line based segmentation
CN115508833A (en) GNSS BI-SAR river boundary detection system
CN115077406A (en) InSAR deformation monitoring method using bridge target extraction result
CN114879180A (en) Seamless situation perception method for real-time fusion of unmanned ship-borne multi-element multi-scale radar
Colin et al. Segmentation of rainfall regimes by machine learning on a colocalized Nexrad/Sentinel-1 dataset
CN113552648A (en) Astronomical positioning and automatic tracking method and system for space target and electronic equipment

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