CN101980296A - Spectral de-aliasing-based Yellow River mainstream line detection method - Google Patents

Spectral de-aliasing-based Yellow River mainstream line detection method Download PDF

Info

Publication number
CN101980296A
CN101980296A CN2010102928774A CN201010292877A CN101980296A CN 101980296 A CN101980296 A CN 101980296A CN 2010102928774 A CN2010102928774 A CN 2010102928774A CN 201010292877 A CN201010292877 A CN 201010292877A CN 101980296 A CN101980296 A CN 101980296A
Authority
CN
China
Prior art keywords
river
line
zone
region
point
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN2010102928774A
Other languages
Chinese (zh)
Inventor
张艳宁
佘红伟
赵娜
刘学工
段峰
张海超
杨旭普
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.)
Northwestern Polytechnical University
Original Assignee
Northwestern Polytechnical 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 Northwestern Polytechnical University filed Critical Northwestern Polytechnical University
Priority to CN2010102928774A priority Critical patent/CN101980296A/en
Publication of CN101980296A publication Critical patent/CN101980296A/en
Pending legal-status Critical Current

Links

Images

Landscapes

  • Image Analysis (AREA)

Abstract

The invention discloses a spectral de-aliasing-based Yellow River mainstream line detection method, which is used for solving the technical problem of low accuracy in the conventional Yellow River mainstream line detection method. The technical scheme is that: the method comprises the following steps of: roughly segmenting the river according to spectral characteristics of a water area in a remote sensing image; extracting bank lines by utilizing edge detection and the like to determine a river area; segmenting the river according to the curve degree and direction of the river reach and the river distribution between south and north water edges; for the segmented riverway, extracting area de-aliasing-based mainstream points by utilizing the spectral and physical characteristics; and connecting the extracted mainstream points by adopting a multiscale analysis method to form a final mainstream line. The accuracy of the Yellow River mainstream line detection method is improved.

Description

Separate the main slide line detecting method in mixed the Yellow River based on spectrum
Technical field
The present invention relates to the main slide line detecting method in a kind of the Yellow River, particularly separate the main slide line detecting method in mixed the Yellow River based on spectrum.
Background technology
Being extracted in of the main slide line in the Yellow River all is by artificial drafting traditionally, and this method is not only time-consuming, effort, and is subjected to the influence of natural conditions such as weather easily, the more important thing is, is difficult in time grasp in flood season the main situation of change of slipping.In recent years, remote sensing image has been widely used in fields such as water body identification, river extraction, water quality detection, big flood detection, waters change-detection, and methods that these researchs are adopted are the characteristic that is shown on image according to water body basically and use some traditional sorting techniques to detect.The continuous development that remote sensing technology and remote sensing image are handled, making the utilization remote sensing image carry out the main slide line detection in the Yellow River becomes possibility.The problem that the main slide in river course shows in husky matter river course of the changeable alluvial of river gesture is outstanding, because less to the gesture research of river, heavily silt-carrying river river course in the world, so, also do not report at present about the applied research achievement of main slide in remote sensing image decipher river course and river gesture figure.Domestic, Preliminary Applications has only been carried out in river course main slide remote sensing image decipher at present in the Yellow River, yet, image data is handled and decipher relies on artificial visual, whole process is time-consuming, simultaneously, also has the difficulty that is difficult to accurately extract the main slide in river course according to the visual interpretation of remote sensing image.
Summary of the invention
In order to overcome the deficiency of existing the Yellow River main slide line detecting method poor accuracy, the invention provides and a kind ofly separate the main slide line detecting method in mixed the Yellow River based on spectrum, this method is carried out the river coarse segmentation according to the spectrum characteristic in waters in the remote sensing images; Utilize rim detection etc. to extract shoreline again, determine river region; Degree of crook and the distribution of the river between direction and the north and south flowage line according to the section are carried out segmentation to the river; To the river course after the segmentation, utilize main spectrum, physical characteristics of slipping line to extract and separate the mixed main point that slips based on the zone; Use multiscale analysis method, the main point that slips that extracts is connected, form the final main line that slips, can improve the Yellow River main slide line detecting method accuracy.
The technical solution adopted for the present invention to solve the technical problems: a kind ofly separate the main slide line detecting method in mixed the Yellow River, be characterized in comprising the steps: based on spectrum
(a) adopt supervised classification and matching process to carry out the river and cut apart, and sorted image is carried out the morphology processing, merge the zonule, eliminate less beach and bridge according to the features of shape of section, the Yellow River;
(b) adopt the Canny operator that the edge is detected, connection is followed the tracks of at detected edge, connect preliminary north and south, the Yellow River bank flowage line, remove and disturb line segment, obtain one group of useful line segment according to the neighborhood method; Again according to statistical property, the number that the point of being judged in every line segment belongs to a certain bank is greater than a certain threshold value, north and south bank image; The line segment of two sides, north and south is stored in the matrix respectively in certain sequence in order, obtain the complete flowage line of north and south bank;
(c) the dam bank that calculates the space by the curvature of windowing transforms to the curvature territory, and the curved place, position of pushing up of bend is represented in the maximum point position of gained curvature sequence; The transition and linkage point position between two continuous river bends is represented in minimum point position between two continuous threshold points, and the Yellow River is divided into typical section and atypia section;
(d) adopt the SMACC algorithm to carry out pure end member and extract, adopt the AMGS algorithm to seek limit, the picture element position of brightness maximum is kept among the q (1)
w 1 = h q ( 1 ) 0
w n = h q ( n ) n - 1
Secondly, concentrate the mapping of removing current selection vector from data with existing
h j n = h j n - 1 - w n F n , j n
F n , j n = α n , j ( w n , h j n - 1 ) ( w n , w n )
Repeat above two steps, satisfy certain threshold value up to residual error; Determine the position of main slide point;
(e) 1. with the multiple dimensioned geometric areas that is divided into of original data space; Out to out
Figure BSA00000284737800025
The definition geometric areas is that (j, k l) are the parallelogram zone to R, and its horizontal width is w=2 -j, vertical thickness is t=2 J-J+rJ=1...J wherein; Definition k is the position of region R level; Defining variable L represents the inclined degree in zone; H is the upright position of region R;
Two-dimensional space is carried out two along transverse axis to be advanced to divide;
Figure BSA00000284737800026
Be the out to out of dividing; The width of each longitudinal region of dividing is w=2 -j, j=1...J wherein; Any 2 connections on each border, the longitudinal region left and right sides are being done two parallel lines up and down with this line respectively as center line, obtain a parallelogram zone, and establishing thickness is t=2 J-J+rThis parallelogram zone is a data structure; Slope with s definition center line; Upright position with h definition center line left end point; Define two scale factors, δ 1=t/ (Vw), δ 2=t/U, the resolution of corresponding slope and upright position respectively; Data space is divided the set in parallelogram zone under each yardstick; The absolute value of the slope that defined range tilts is no more than S; With the parallelogram region representation is that (l i), wherein is respectively yardstick, 0≤k＜1/w to R for j, k -1,-S δ 1≤ l≤S δ 1, 0≤i≤δ 2 -1-1; For R (j, k, l, i), the horizontal ordinate on left vertical limit is x=kw, the intersection point on center line and left vertical limit is y=i δ 2, slope s=l δ 1
Two parallelogram region R 1(j 1, k 1, l 1, i 1) and region R 2(j 2, k 2, l 2, i 2) between the condition of continuity:
● two zones under same yardstick, i.e. j 1=j 2
● two zones are adjacent, promptly | and k 1-k 2|=1;
● the intersection point distance in public vertical sideline and two zones is very near, promptly | and l 1+ i 1-l 2|＜v;
● the middle line slope in two zones is more or less the same, promptly | and l 1-l 2|＜u;
With two zone definitions that satisfy above-mentioned four conditions is the good zone of continuity;
2. add up counting in the geometric areas under each yardstick, and select significant zone according to threshold value;
3. under each yardstick, set up a non-directed graph G j=(V j, E j), the geometric areas when Count＞N is as the vertex v ∈ V of figure jIn the formula, Count is counting in the region R, and N is the definition threshold value in the region R;
If two geometric areas satisfy successional condition, just between these two summits, connect a limit e ∈ E jThe quantity of the point in the computational geometry zone is is one by one accepted or rejected and the continuity relation is set up the limit according to threshold value N, obtains non-directed graph;
4. use the depth-first search algorithm in each non-directed graph of setting up, search longest path, the longest path that searches out are the main line that slips.
The invention has the beneficial effects as follows: the spectrum characteristic owing to according to waters in the remote sensing images, carry out the river coarse segmentation; Utilize rim detection etc. to extract shoreline again, determine river region; Degree of crook and the distribution of the river between direction and the north and south flowage line according to the section are carried out segmentation to the river; To the river course after the segmentation, utilize main spectrum, physical characteristics of slipping line to extract and separate the mixed main point that slips based on the zone; Use multiscale analysis method, the main point that slips that extracts is connected, form the final main line that slips, improved the Yellow River main slide line detecting method accuracy.
Below in conjunction with the drawings and specific embodiments the present invention is elaborated.
Description of drawings
Accompanying drawing is to the present invention is based on spectrum to separate mixed the Yellow River main slide line detecting method process flow diagram.
Embodiment
With reference to accompanying drawing.1, the river is cut apart.
Importing a secondary TM multispectral image, at first is the river coarse segmentation.Adopt spectral classification and matching technique, as: spectrum to flux matched, mahalanobis distance is cut apart and method such as Gauss Markov is carried out the spectrum picture classification, and according to the aftertreatment of classifying of the features of shape of section, the Yellow River, merge as: zone etc.Divide time-like to adopt supervised classification method, because image is subjected to the influence of factors such as weather, in the image there be than big-difference the spectrum of the upper reaches of the Yellow River and downstream water body, have very mistake if adopt single sample to carry out the classification of spectrum angle, therefore adopt two kinds of samples respectively image to be classified, then composograph.Because the influence of bridge and beach is arranged in the image, be not easy to the extraction of two sides, north and south flowage line, under the situation that requires to be fit to further study in edge definition image has been carried out the morphology processing, promptly image expands and corrodes operation, has eliminated less beach and bridge.
2, bank line extracts.
Adopt the Canny operator that the edge is detected, and connection is followed the tracks of at detected edge, sort out the flowage line of north and south, river bank, for next step river segmentation is got ready.For the ease of extracting the two sides, north and south, need resulting edge line is carried out connection tracking; According to the neighborhood method, utilize a border following algorithm to arrive and obtain one group of continuous segments, remove some short interference line segments according to length, can remove beach, thereby obtain one group of useful line segment according to the style characteristic of beach simultaneously.According to the distribution characteristics of the Yellow River water body, observe it and have certain horizontal and vertical features, but algorithm for design carries out the judgement of north and south bank to every section line segment in view of the above.Only to need to consider to obtain the information of which bank under this whole section flowage line to the judgement that some representational point in every section carries out the north and south bank based on efficiency of algorithm.According to statistical property, the point of being judged in every section belongs to the number of a certain bank can judge that greater than a certain threshold value this section is to belong to southern bank or northern bank, otherwise, then belong to an other bank, thereby obtain north and south bank image.As required the line segment of two sides, north and south is stored in the matrix respectively in certain sequence in order, thereby obtain the complete flowage line in two sides, be beneficial to next step the main line drawing that slips.
3, river segmentation.
With regard to the Yellow River main slide line problem the Yellow River is divided into typical section and atypia section.The typical case section is divided into straight little curved, the crooked and branch branch of a river three classes again.
The shape of each class section all has very big difference, and the main mode one of describing difference is to utilize tortuosity factor and curvature to describe the degree of crook and the direction of section, and the 2nd, distribute to determine whether to exist by the river between the flowage line of north and south and divide a branch of a river.Space relationship between section and the section also is one of main slide line foundation in the Yellow River of judging, therefore, it also is very important calculating for the relation of the space distribution between the section, the present invention is sections such as Curved Continuous is joined, bend is come over and pledged allegiance to the space distribution contextual definition, utilizes the continuous variation arrangement between the section to be distinguished from the space.Concrete segmentation method is as follows:
A) the dam bank that calculates the space by the curvature of windowing transforms to the curvature territory, and the curved place, position of pushing up of bend has been represented in the maximum point position of gained curvature sequence.
B) the transition and linkage point position between two continuous river bends is represented in the minimum point position between two continuous threshold points.
4, the main point that slips extracts.
Separate the mixed main extracting method that slips based on the zone.
Background outside the shielding river course is carried out the PCA conversion to certain section, and the regional area of getting after the conversion is separated mixed.(setting the end member number is four for sequential maximum angle convex cone, SMACC) algorithm, carries out pure end member and extracts: at first, adopt limit to determine convex cone, and define first end member with this to adopt one by one the maximum angular convex cone.Afterwards, in existing cone, use the oblique projection of satisfying above-mentioned constraint condition and generate next end member.Increase cone and generate new end member.This method repeats always, satisfies existing end member in certain tolerance, the convex cone until generating one, perhaps until the end member classification number that has satisfied appointment.
When determining limit, adopt successive orthogonalization algorithm AMGS (adaptation of the augmented modifiedGram Schmidt) to seek limit.At first, from H, choose the vector that will keep.Initial vector can picked at random, also can choose according to certain criterion, and here, what choose is the picture element of brightness maximum, and its position is kept among the q (1).
w 1 = h q ( 1 ) 0
w n = h q ( n ) n - 1
Secondly, concentrate the mapping of removing current selection vector from data with existing.
h j n = h j n - 1 - w n F n , j n
F n , j n = α n , j ( w n , h j n - 1 ) ( w n , w n )
Repeat above two steps, satisfy certain threshold value up to residual error.
When determining the abundance image, every kind of material is represented with a subspace of containing its variation,, obtained each content of material in the end member by to this sub spaces upslide shadow.Utilize the limit vector not represent, but not the then passable characteristic of limit is sought the subspace that its each end member covers, thereby determine abundance by the non-negative wire combination of other point in the image.In the abundance image of various end member correspondences, choose the abundance image of end member spectrum correspondence with monotone increasing characteristics.Determine after the abundance image, the location that utilizes the maximum abundance localization method of section to lead the slide point, promptly, each section (river cross-section is approximate with perpendicular line) pixel value to the river course carries out extreme value analysis, find out extreme point, be the point of brightness maximum in the abundance image, thereby determine the position of main slide point.
5, the main point that slips connects.
Above main the slide in the method that point extracts, the main point that slips that obtains need be connected and could obtain line segment.The present invention adopts the curve method of attachment of multiscale analysis:
At first will lead slide point data space and under each yardstick, divide, to each zone that marks off, counting in the statistical regions.Make up non-directed graph then.The longest path of last searching in the drawings.Concrete step is:
A) multiple dimensioned division: original data space is carried out multiple dimensioned division, be divided into geometric areas.At first definition has the multiple dimensioned geometric areas of different directions, yardstick, length.Out to out
Figure BSA00000284737800055
The definition geometric areas is that (j, k l) are the parallelogram zone to R, and its horizontal width is w=2 -j, vertical thickness is t=2 J-J+rJ=1...J wherein.Width and thickness depend on the number of the size and the data point of yardstick.Definition k is the position of region R level.Defining variable L represents the inclined degree in zone in addition.H is the upright position of region R.
Then two-dimensional space being carried out two along transverse axis advances to divide.
Figure BSA00000284737800056
Be the out to out of dividing.The width of each longitudinal region of dividing is w=2 -j, j=1...J wherein.Any 2 connections on each border, the longitudinal region left and right sides are being done two parallel lines up and down with this line respectively as center line, obtain a parallelogram zone, and establishing thickness is t=2 J-J+rThis parallelogram zone is our needed data structure, obtains more parallelogram zone by translation and rotation center line.Use the slope of s definition center line.And the upright position of h definition center line left end point.Define two scale factors in addition, δ 1=t/ (Vw), δ 2=t/U, the resolution of corresponding slope and upright position respectively.Data space is divided the set in parallelogram zone under each yardstick.The absolute value of the slope that defined range tilts is no more than S.With the parallelogram region representation is that (l i), wherein is respectively yardstick, 0≤k＜1/w to R for j, k -1,-S δ 1≤ l≤S δ 1, 0≤i≤δ 2 -1-1.Therefore for R (j, k, l, i), the horizontal ordinate on left vertical limit is x=kw, the intersection point on center line and left vertical limit is y=i δ 2, slope s=l δ 1
Define two parallelogram region R 1(j 1, k 1, l 1, i 1) and region R 2(j 2, k 2, l 2, i 2) between the condition of continuity.
● two zones under same yardstick, i.e. j 1=j 2
● two zones are adjacent, promptly | and k 1-k 2|=1.
● the intersection point distance in public vertical sideline and two zones is very near, promptly | and l 1+ i 1-l 2|＜v.
● the middle line slope in two zones is more or less the same, promptly | and l 1-l 2|＜u.
With two zone definitions that satisfy above-mentioned four conditions is the good zone of continuity.
B) data statistics: add up counting in the geometric areas under each yardstick, and select significant zone according to threshold value.
C) structure non-directed graph: the continuity of utilizing marking area and zone to see makes up non-directed graph.
Under each yardstick, set up a non-directed graph G j=(V j, E j), use the vertex v ∈ V of each above-mentioned geometric areas as figure j, but be not all to set up a summit for each geometric areas.But to choose wherein satisfactoryly, be called the candidate region.Choosing of candidate region is quantity decision by data point in the zone.If counting in the region R is Count, definition threshold value N, when Count＞N, this zone is the candidate region.
If two geometric areas satisfy successional condition, just between these two summits, connect a limit e ∈ E jThe quantity of the point in the computational geometry zone one by one, N accepts or rejects according to threshold value, sets up the limit according to the continuity relation then, thereby sets up a non-directed graph.
The search longest path: use the depth-first search algorithm in each non-directed graph of setting up, search longest path, the longest path that searches out are the main line that slips.

Claims (1)

1. separate the main slide line detecting method in mixed the Yellow River based on spectrum for one kind, it is characterized in that comprising the steps:
(a) adopt supervised classification and matching process to carry out the river and cut apart, and sorted image is carried out the morphology processing, merge the zonule, eliminate less beach and bridge according to the features of shape of section, the Yellow River;
(b) adopt the Canny operator that the edge is detected, connection is followed the tracks of at detected edge, connect preliminary north and south, the Yellow River bank flowage line, remove and disturb line segment, obtain one group of useful line segment according to the neighborhood method; Again according to statistical property, the number that the point of being judged in every line segment belongs to a certain bank is greater than a certain threshold value, north and south bank image; The line segment of two sides, north and south is stored in the matrix respectively in certain sequence in order, obtain the complete flowage line of north and south bank;
(c) the dam bank that calculates the space by the curvature of windowing transforms to the curvature territory, and the curved place, position of pushing up of bend is represented in the maximum point position of gained curvature sequence; The transition and linkage point position between two continuous river bends is represented in minimum point position between two continuous threshold points, and the Yellow River is divided into typical section and atypia section;
(d) adopt the SMACC algorithm to carry out pure end member and extract, adopt the AMGS algorithm to seek limit, the picture element position of brightness maximum is kept among the q (1)
w 1 = h q ( 1 ) 0
w n = h q ( n ) n - 1
Secondly, concentrate the mapping of removing current selection vector from data with existing
h j n = h j n - 1 - w n F n , j n
F n , j n = α n , j ( w n , h j n - 1 ) ( w n , w n )
Repeat above two steps, satisfy certain threshold value up to residual error; Determine the position of main slide point;
(e) 1. with the multiple dimensioned geometric areas that is divided into of original data space; Out to out
Figure FSA00000284737700015
The definition geometric areas is that (j, k l) are the parallelogram zone to R, and its horizontal width is w=2 -j, vertical thickness is t=2 J-J+rJ=1 wherein ... J; Definition k is the position of region R level; Defining variable L represents the inclined degree in zone; H is the upright position of region R;
Two-dimensional space is carried out two along transverse axis to be advanced to divide; Be the out to out of dividing; The width of each longitudinal region of dividing is w=2 -j, j=1 wherein ... J; Any 2 connections on each border, the longitudinal region left and right sides are being done two parallel lines up and down with this line respectively as center line, obtain a parallelogram zone, and establishing thickness is t=2 J-J+rThis parallelogram zone is a data structure; Slope with s definition center line; Upright position with h definition center line left end point; Define two scale factors, δ 1=t/ (Vw), δ 2=t/U, the resolution of corresponding slope and upright position respectively; Data space is divided the set in parallelogram zone under each yardstick; The absolute value of the slope that defined range tilts is no more than S; With the parallelogram region representation is that (l i), wherein is respectively yardstick, 0≤k＜1/w to R for j, k -1,-S δ 1≤ l≤S δ 1, 0≤i≤δ 2 -1-1; For R (j, k, l, i), the horizontal ordinate on left vertical limit is x=kw, the intersection point on center line and left vertical limit is y=i δ 2, slope s=l δ 1
Two parallelogram region R 1(j 1, k 1, l 1, i 1) and region R 2(j 2, k 2, l 2, i 2) between the condition of continuity:
● two zones under same yardstick, i.e. j 1=j 2
● two zones are adjacent, promptly | and k 1-k 2|=1;
● the intersection point distance in public vertical sideline and two zones is very near, promptly | and l 1+ i 1-l 2|＜v;
● the middle line slope in two zones is more or less the same, promptly | and l 1-l 2|＜u;
With two zone definitions that satisfy above-mentioned four conditions is the good zone of continuity;
2. add up counting in the geometric areas under each yardstick, and select significant zone according to threshold value;
3. under each yardstick, set up a non-directed graph G j=(V j, E j), the geometric areas when Count＞N is as the vertex v ∈ V of figure jIn the formula, Count is counting in the region R, and N is the definition threshold value in the region R;
If two geometric areas satisfy successional condition, just between these two summits, connect a limit e ∈ E jThe quantity of the point in the computational geometry zone is is one by one accepted or rejected and the continuity relation is set up the limit according to threshold value N, obtains non-directed graph;
4. use the depth-first search algorithm in each non-directed graph of setting up, search longest path, the longest path that searches out are the main line that slips.
CN2010102928774A 2010-09-25 2010-09-25 Spectral de-aliasing-based Yellow River mainstream line detection method Pending CN101980296A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN2010102928774A CN101980296A (en) 2010-09-25 2010-09-25 Spectral de-aliasing-based Yellow River mainstream line detection method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN2010102928774A CN101980296A (en) 2010-09-25 2010-09-25 Spectral de-aliasing-based Yellow River mainstream line detection method

Publications (1)

Publication Number Publication Date
CN101980296A true CN101980296A (en) 2011-02-23

Family

ID=43600797

Family Applications (1)

Application Number Title Priority Date Filing Date
CN2010102928774A Pending CN101980296A (en) 2010-09-25 2010-09-25 Spectral de-aliasing-based Yellow River mainstream line detection method

Country Status (1)

Country Link
CN (1) CN101980296A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103123723A (en) * 2013-01-23 2013-05-29 中国人民解放军信息工程大学 Flowage line extracting method based on Canny edge detection and active contour model
CN105719300A (en) * 2016-01-22 2016-06-29 黄河水利委员会信息中心 Riverway main stream line detection method based on SNE manifold learning
CN105913023A (en) * 2016-04-12 2016-08-31 西北工业大学 Cooperated detecting method for ice of The Yellow River based on multispectral image and SAR image
CN105913430A (en) * 2016-04-12 2016-08-31 西北工业大学 Cooperated extracting method for main ice information of The Yellow River based on multispectral remote sensing image

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103123723A (en) * 2013-01-23 2013-05-29 中国人民解放军信息工程大学 Flowage line extracting method based on Canny edge detection and active contour model
CN105719300A (en) * 2016-01-22 2016-06-29 黄河水利委员会信息中心 Riverway main stream line detection method based on SNE manifold learning
CN105719300B (en) * 2016-01-22 2018-11-13 黄河水利委员会信息中心 River based on SNE manifold learnings is main to slip line detecting method
CN105913023A (en) * 2016-04-12 2016-08-31 西北工业大学 Cooperated detecting method for ice of The Yellow River based on multispectral image and SAR image
CN105913430A (en) * 2016-04-12 2016-08-31 西北工业大学 Cooperated extracting method for main ice information of The Yellow River based on multispectral remote sensing image
CN105913430B (en) * 2016-04-12 2019-01-15 西北工业大学 Yellow River Main based on multi-spectral remote sensing image slips information synergism extracting method
CN105913023B (en) * 2016-04-12 2019-06-21 西北工业大学 The Yellow River ice slush collaborative detection method based on multispectral image and SAR image

Similar Documents

Publication Publication Date Title
Wei et al. Toward automatic building footprint delineation from aerial images using CNN and regularization
CN102622607B (en) Remote sensing image classification method based on multi-feature fusion
CN103049763B (en) Context-constraint-based target identification method
CN107392141B (en) Airport extraction method based on significance detection and LSD (least squares distortion) line detection
Mason et al. Extraction of tidal channel networks from airborne scanning laser altimetry
CN102750703B (en) Remote sensing image airport automatic detecting method based on linear cutting
CN102663382B (en) Video image character recognition method based on submesh characteristic adaptive weighting
CN101145196B (en) Quick fingerprint identification method based on strange topology structure
CN102496034A (en) High-spatial resolution remote-sensing image bag-of-word classification method based on linear words
CN101980295A (en) Method for detecting mainstream line of Yellow River based on skewness analysis
CN102043958A (en) High-definition remote sensing image multi-class target detection and identification method
CN112184746A (en) Transmission line insulator defect analysis method
CN103473551A (en) Station logo recognition method and system based on SIFT operators
CN103383733A (en) Lane video detection method based on half-machine study
Gamba et al. Urban climate zone detection and discrimination using object-based analysis of VHR scenes
CN105931241A (en) Automatic marking method for natural scene image
CN101980296A (en) Spectral de-aliasing-based Yellow River mainstream line detection method
CN105184225A (en) Multinational paper money image identification method and apparatus
CN106127144B (en) Using when compose the point source risk source extraction method of empty integral feature model
CN102013097A (en) Yellow River main humping line detection method based on spectrum similarity and space continuity
CN101840513A (en) Method for extracting image shape characteristics
CN110084107A (en) A kind of high-resolution remote sensing image method for extracting roads and device based on improvement MRF
CN107146233A (en) Granulometry Segmentation based on petrographic thin section polarisation sequence chart
Ijaz et al. Detection of hydromorphologic characteristics of Indus River Estuary, Pakistan, using satellite and field data
CN109543498A (en) A kind of method for detecting lane lines based on multitask network

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C02 Deemed withdrawal of patent application after publication (patent law 2001)
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20110223