CN101510310A - Method for segmentation of high resolution remote sensing image based on veins clustering constrain - Google Patents
Method for segmentation of high resolution remote sensing image based on veins clustering constrain Download PDFInfo
- Publication number
- CN101510310A CN101510310A CNA2009100463483A CN200910046348A CN101510310A CN 101510310 A CN101510310 A CN 101510310A CN A2009100463483 A CNA2009100463483 A CN A2009100463483A CN 200910046348 A CN200910046348 A CN 200910046348A CN 101510310 A CN101510310 A CN 101510310A
- Authority
- CN
- China
- Prior art keywords
- zone
- texture
- clustering
- segmentation
- optimum
- 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
Links
Landscapes
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a high-resolution remote sensing image segmentation method based on texture clustering constraint in the field of remote sensing technology, which comprises the concrete steps of: step 1, calculating Gabor energy textures of all regions in an image, carrying out clustering of all the regions by using FCM according to texture distance, and setting a texture clustering tag for each region according to clustering results; step 2, setting up a comprehensive distance space model by using features such as spectrum, texture, shape and the like, and adding texture clustering distance to restrict and lead region merging to be performed along the homogeneous texture direction; and step 3, setting up an RAG model and an NNG model according to the comprehensive distance, and merging the regions according to global optimum. The real regional boundary is obtained by the interaction of the texture clustering and optimal merging sequence during the merging. The method can better segment the texture region in the high-resolution image and improve the overall segmentation accuracy of the image.
Description
Technical field
What the present invention relates to is the image partition method in a kind of remote sensing technology field, specifically is a kind of method for segmentation of high resolution remote sensing image based on veins clustering constrain.
Background technology
Along with improving constantly of satellite spatial resolution, OO image analysis method begins in the widespread use of remote sensing image processing field.With respect to the sorting technique based on pixel, OO image analysis method can reduce The noise greatly, extracts more feature, and is easy to and GIS (GeographicalInformation System, Geographic Information System) combines.But the key component in the object-oriented image analysis method---image segmentation---is not well solved all the time.
According to feature space, image partition method can be divided three classes substantially: based on the dividing method of spectrum, based on the dividing method of texture and based on spectrum and texture will in conjunction with dividing method, wherein the method for cutting apart based on spectrum is a most frequently used method of present stage.The Definiens Developer Software company of Germany has proposed a kind of remote sensing image segmentation method based on spectrum, promptly seek spectrum homogeney zone in the remote sensing images, be specially: change according to the spectrum variance that merges rear region and judge interregional spectral similarity degree to be combined, then according to the spectral similarity degree, successively the zone with higher similarity is merged, when the similarity between All Ranges stops during all greater than threshold value merging, obtain segmentation result.This method has obtained application in the software Definiens of the said firm.
Find by prior art documents, people such as U.Benz are in " ISPRS Journal ofPhotoGrammetry and Remote Sensing " (the photogrammetric and remote sensing of ISPRS), Volume.58, provided detailed introduction in " Multi-resolution, the object-oriented fuzzy analysisof remote sensing data for GIS-ready information " that delivers on the page.239-258 (using remote sensing images to extract the multiresolution object-oriented fuzzy analysis method of the GIS information needed) literary composition.But in high-resolution remote sensing image, more zone is a texture region, its homogeneity not on spectrum.Therefore, when the spectrum dividing method in the use Definiens software was cut apart high-resolution remote sensing image, a lot of texture regions was divided into broken fritter, can not obtain complete zone, and segmentation precision is lower.
The method of using spectrum and texture to combine is one of method that improves the high-resolution remote sensing image segmentation precision.People such as Y.Deng are in " IEEE Trans.Pattern Anal.Mach.Intell. " (IEEE pattern-recognition and machine function), Volume.14,2001, the JSEG method that a kind of spectrum and texture combine is proposed in " Unsupervisedsegmentation of spectral-texture regions in images and video " (based on the non-supervision dividing method of spectrum texture region in image and the video) literary composition of delivering on the page.800-810, be specially: spectrum is quantized, and obtain spatial texture J image according to the frequency that the part quantizes spectrum, the result who uses region growing method to cut apart to the end to the J image then.This method has obtained result preferably when scene image is cut apart, but high-resolution remote sensing image content complexity quantizes meeting reduction amount of image information to spectrum, and the texture description that obtains is inaccurate, the also difficult real border that accurately obtains atural object.And the JSEG method does not realize the combination semantically of spectrum and texture with texture and spectrum separate processes.How to use various features to cut apart high-resolution remote sensing image, and obtain that the accurate border of texture region remains one of present stage open question in the image.
Summary of the invention
The object of the invention is to overcome the defective that prior art middle high-resolution remote sensing images are cut apart middle texture information underutilization, a kind of method for segmentation of high resolution remote sensing image based on veins clustering constrain is proposed, cut apart the order that time domain merges by veins clustering pre-service influence, the zone is merged according to the direction of texture homogeneity, and the accurate border in zone has been found in the interaction of using optimum collating sequence and veins clustering in merging process, has improved the precision of cutting apart.
The present invention is achieved by the following technical solutions, and it is as follows to the present invention includes step:
The first step, obtain zone-texture cluster label: the Gabor energy textures of All Ranges (high wave energy texture) in the computed image, and use FCM (fuzzy C-means clustering algorithm) that All Ranges is carried out cluster according to texture, set veins clustering label for each zone according to cluster result;
Second goes on foot, and sets up the metric space model of semantic congruence: use features such as spectrum, texture and shape to set up the comprehensive distance spatial model, and add veins clustering apart from the zone is merged into row constraint, merging can be carried out along the direction of texture homogeneity;
In the 3rd step, use the graph model algorithm that the zone is merged fast: set up RAG (regional connection layout) and NNG (arest neighbors map interlinking) model according to comprehensive distance, and according to global optimum to the zone to merging, obtain final segmentation result.In merging process, obtain the real border in zone by the interaction of veins clustering and optimum collating sequence.
Described zoning Gabor energy textures, and use FCM to region clustering, obtain zone-texture cluster label, be specially: use 3 yardsticks, 8 direction Gabor filtering that image is carried out convolution, get the count root of quadratic sum of corresponding symmetry and antisymmetry Gabor filtering result and obtain the Gabor energy, the Gabor energy is 24 dimensional vectors, uses this vectorial Euclidean distance to carry out cluster and obtains zone-texture cluster label.
Described semantic congruence metric space is specially: use the veins clustering label that regional merge order is retrained the zone merging is carried out according to texture region homogeneity direction.Comprehensively use information such as spectrum, texture and shape then, and add the veins clustering distance, set up the metric space model that uses many features.In this metric space model, during interregional texture homogeneity, spectrum intervals is less; During the spectrum homogeneity, texture is less.
Described quick merging method based on RAG and NNG figure is specially: use RAG figure to describe syntople between the zone, and the combined distance between posting field; Use NNG figure to describe the corresponding optimum in each zone and merge the zone, and write down optimum combined distance.In NNG figure, the optimum each other zone that merges the zone is right to be that local optimum in the image merges the zone, and all local optimum zones are obtained global optimum's collating sequence to ordering, can obtain segmentation result according to this sequence merging.In merging process, the zone-texture cluster label after the merging changes, and then influences the merge order of sequence, can obtain the real border of atural object by this interaction.
This drawing method has been accelerated the efficient that merges greatly, has reduced the time complexity of partitioning algorithm.
The present invention can obtain the segmentation result of high-resolution remote sensing image preferably, and texture region is comparatively complete among the result, and the border is more accurate, and segmentation precision is higher.
The present invention is incorporated into veins clustering in the image segmentation algorithm, has solved emphatically how to use various features accurately to cut apart the problem of high-resolution remote sensing image.In partitioning algorithm, taken into full account the relation between spectrum homogeneous region and the texture homogeneous region, the texture region in the image can be split exactly.In addition, algorithm has obtained the accurate border in zone by the interaction of veins clustering and optimal region merging.This algorithm has been avoided the defective in traditional remote sensing image segmentation method: use simple light spectrum signature difficulty is partitioned into the correct texture zone.Therefore, compare with classic method, the present invention has higher segmentation precision.The partitioning algorithm that the present invention proposes based on veins clustering constrain, informixs such as texture information and spectrum are used, significantly improved the segmentation precision of high-resolution remote sensing image, puted forth effort to solve the problem that the remote sensing image process field uses many features to cut apart, subject development has been had impetus.
Embodiment
Below embodiments of the invention are elaborated: present embodiment is being to implement under the prerequisite with the technical solution of the present invention, provided detailed embodiment and concrete operating process, but protection scope of the present invention is not limited to following embodiment.
Present embodiment uses the FCM algorithm that cluster is carried out in the zone according to textural characteristics, and gives label with cluster result to the zone.Many characteristic distances model of semantic congruence is proposed then.Distance model comprehensively uses the interregional homogeney distances of measure information such as spectrum, texture and shape, and adds interregional veins clustering apart from regional merging can be carried out according to texture homogeneity direction.At last, many characteristic distances model is applied in RAG and the NNG graph model, obtains segmentation result by the merging of quick global optimal region.In merging process, the interaction between veins clustering and optimal characteristics merge has guaranteed the accuracy on texture region border.
1, obtain regional Gabor energy textures feature, be implemented as follows:
The Gabor function of one 2 dimension can be written as following form:
x′＝xcosθ+ysinθ
y′＝-xsinθ+ycosθ
Wherein λ is a wavelength, and 1/ λ can be used for representing the high and low frequency center of frequency domain internal object; θ is the strip direction of Gabor; R is a length breadth ratio, has determined the index ellipsoid of Gabor; δ/λ is the frequency domain bandwidth of Gabor wave filter.
Be the phase parameter of Gabor wave filter, when phase place is 0 and during π, wave filter is the center symmetry; When phase place was 1/2 π and-1/2 π, wave filter was an antisymmetric function.
The Gabor energy function can be defined as following form:
r
λ, θ, 0(x, y) and
Be to use respectively
With
To the result after the image convolution.
2, set up the metric space model of semantic congruence, specific implementation is:
Use the metric space model of information such as spectrum, texture and shape to be:
d＝d
t·d
c·d
s (3)
Wherein, d
tBe texture, d
cBe color distance, d
sBe the shape distance.
Texture d
tBe defined as:
d
t＝d
tr+d
tc (4)
D wherein
TrBe interregional Gabor texture, it is defined as:
B is the wave band number; u
1And u
2Be respectively the Gabor average energy in zone 1 and zone 2.
d
TcBe interregional veins clustering distance, it is defined as:
e
iAnd e
jBe respectively the center Gabor energy of zone 1 and zone 2 affiliated cluster classification i and j.
Color distance d
cDefinition:
B is the wave band number; w
iWeight for each wave band of image; n
mBe the pixel count after zone 1 and zone 2 merge, n
1, n
2Be respectively the pixel count in zone 1 and zone 2; σ is the mean square deviation of object under the i wave band, and its subscript represents to merge rear region, zone 1 and zone 2 respectively.
Shape is apart from d
sBe defined as:
L wherein
mBe the length of side to picture, n
cBe interregional adjacent pixels number, d
cBe interregional spectrum intervals, n
mBe the area pixel number after merging.
3, set up RAG and NNG graph model, specific implementation is:
RAG represents the syntople of region junction, and it is described below: a width of cloth K the zone image can be expressed as simple graph G=(V, E), V is the zone of image, is the vertex set V={1 of figure in graph structure, 2 ..., K}; E is the regional adjacent annexation of image, is the limit collection in graph structure
。When the zone is adjacent, have limit e between region junction, e represents interregional merging cost, can draw by formula (3); When the zone is non-conterminous, there is not the limit to exist between the regional summit, promptly merge cost for infinitely great.
The merge order of NNG posting field, it is described as: establish the image in a width of cloth K zone, its RAG figure is that (V, E), then G has subgraph G to G=
m=(V
m, E
m), V wherein
m=V; E
m={ e
i=mine
I, j| i, j ∈ (1,2 ... k) }, i.e. the limit collection E of subgraph
mFor and vertex v
iThe minimum weights limit e that is associated
I, jSet.For non-directed graph G
m, make its limit collection E
mIn each element increase a direction and obtain arc collection A
m={ a
i=Arc (v
i, v
j) | e
I, j∈ E
m, non-directed graph G then
mGenerate its oriented graph G
m=(V
m, A
m), promptly NNG schemes.Each vertex v
iIn-degree differ and be decided to be 1, but out-degree all is 1, and with v
iFor the head of the arc of tail promptly is the zone of its homogeneity.Work as A
mIn have ring collection A
c={ a
i, a
j| a
i∈ A
m, a
j∈ A
mThe time, v
iWith v
jOptimum each other, also be that the local optimum zone is right.
Algorithm is only searched for the ring set and is decided overall collating sequence.For piece image, the rarest ring collection of NNG has at most
Individual ring collection is so NNG greatly reduces the time of search.Use the position of an auxiliary sequencel record ring collection node.Sequence sorted from small to large by interregional merging cost, and what obtain is exactly global optimum's collating sequence.
4, the zone merging obtains segmentation result, and specific implementation is:
Merge according to global optimum's collating sequence, the Gabor energy of average Gabor energy textures as new region used in each back that merges, and calculate new region in the heart distance in each veins clustering according to formula (5), will give new region apart from the veins clustering classification of minimum.Recomputate combined distance between new region and the neighboring region according to formula (1) then, upgrade RAG and NNG, seek new region respective rings subclass, and upgrade optimum collating sequence, continue to merge.When the right combined distance in global optimum zone during, stop to merge and obtain final segmentation result greater than threshold value.
In merging process, change has taken place in the cluster classification that merges rear region, and simultaneously, the region clustering classification after the merging affects the merge order of optimal sequence again.By the interaction that veins clustering and optimal sequence merge, the border of texture region can accurately obtain.
Below further specify the present embodiment applicable cases:
Used the accuracy of two width of cloth view data test partitioning algorithm, a width of cloth comes from the 2.5mSPOT5 true color remote sensing image in Shanghai City, and a width of cloth is the aviation image of Shanghai City Chongming Island 25cm.The SPOT5 image is to obtain on September 30th, 2005, and the intercepting size is that the data of 512 * 512 pixels are with doing experiment.5 class atural objects have mainly been comprised in institute's sampling area: settlement place, industrial land, waters, arable land and fish pond.Wherein, settlement place spectrum is inhomogeneous, but obviously different with other atural object; Industrial land spectrum is inhomogeneous, but presents linear marking; Spectrum is variant between the arable land, but texture is more similar; Waters and fish pond spectrum homogeneity have a fine grain.Aerial image intercepting size is that 512 * 512 data are tested.In aerial image, mainly comprised forest land, river, settlement place, road, booth vegetable plot, irrigated land status such as paddy field and nonirrigated farmland.Wherein, forest land, booth vegetable plot, irrigation paddy field and nonirrigated farmland etc. show visibly different textural characteristics.
With the result manually cut apart as the reference result, the accuracy of two kinds of dividing methods of contrast, a kind of is based on the homogeneous Definiens dividing method of spectrum, another kind is the present embodiment method.Test findings is as follows: (1) SPOT5 image segmentation accuracy is (seeing Table 1) relatively; (2) aerial image is cut apart accuracy relatively (seeing Table 2).
From four aspects the change-detection result is estimated: the segmentation precision of (1) main object: account for the atural object of image major part, as arable land, forest land etc.; (2) segmentation precision of small object: than the atural object of small size but indispensable when image interpretation, as isolated house, distinctive mark etc.; (3) segmentation precision of linear object: atural objects such as river, road; (4) overall segmentation accuracy.
Table 1 SPOT5 image segmentation result accuracy relatively
Table 2 aerial image segmentation result accuracy relatively
From table, can see that for the main object in the image, present embodiment has been obtained 68.2% and 87.5% segmentation precision, far above the segmentation precision of Definiens software 54.5% and 62.5%.This mainly is owing to large tracts of land object clean mark, and present embodiment is by the easier texture region exactly that obtains of veins clustering restriction, and the Definiens method is then than the difficult correct segmentation result that obtains texture region by spectrum.And for small atural object and linear ground object etc., present embodiment has also been obtained the segmentation precision higher than Definiens.From overall accuracy, present embodiment has improved the segmentation precision of 10%-20% with respect to Definiens software.
Visually, present embodiment as industrial building, forest land etc., is obtained good segmentation effect for the atural object with regular veins.These atural objects are cut apart the back structural integrity, and boundary accurate is better than the segmentation result of Definiens far away.
Generally speaking, present embodiment can be obtained the precision higher than classic method for high-resolution remote sensing image, particularly for the regular veins zone in the image.
Claims (10)
1, a kind of method for segmentation of high resolution remote sensing image based on veins clustering constrain is characterized in that, comprises the steps:
The first step, obtain zone-texture cluster label: the Gabor energy textures of All Ranges in the computed image, and use FCM that All Ranges is carried out cluster according to texture, set veins clustering label for each zone according to cluster result;
Second goes on foot, and sets up the metric space model of semantic congruence: use features such as spectrum, texture and shape to set up the comprehensive distance spatial model, and add veins clustering apart from the zone is merged into row constraint;
In the 3rd step, use the graph model algorithm that the zone is merged fast: set up RAG and NNG according to comprehensive distance, and according to global optimum to the zone to merging, obtain net result.
2, the method for segmentation of high resolution remote sensing image based on veins clustering constrain according to claim 1, it is characterized in that, set the veins clustering label, be specially: use 3 yardsticks, 8 direction Gabor filtering that image is carried out convolution, get the count root of quadratic sum of corresponding symmetry and antisymmetry Gabor filtering result and obtain the Gabor energy, the Gabor energy is 24 dimensional vectors, uses this vectorial Euclidean distance to carry out cluster and obtains zone-texture cluster label.
3, the method for segmentation of high resolution remote sensing image based on veins clustering constrain according to claim 1 is characterized in that, described constraint is specially: use the veins clustering label that regional merge order is retrained.
4, the method for segmentation of high resolution remote sensing image based on veins clustering constrain according to claim 1, it is characterized in that, described comprehensive distance spatial model, be specially: comprehensively use spectrum, texture and shape information, and add the veins clustering distance, set up the metric space model that uses many features.
5, the method for segmentation of high resolution remote sensing image based on veins clustering constrain according to claim 1, it is characterized in that, described the zone is merged fast, be specially: use RAG figure to describe syntople between the zone, and the combined distance between posting field; Use NNG figure to describe the corresponding optimum in each zone and merge the zone, and write down optimum combined distance; In NNG figure, the optimum each other zone that merges the zone obtains global optimum collating sequence with all local optimum zones to ordering to being the local optimum in the image, merges according to this sequence to obtain segmentation result.
6, the method for segmentation of high resolution remote sensing image based on veins clustering constrain according to claim 1 is characterized in that, described merging is meant the real border that obtains the zone in merging process by the interaction of veins clustering and optimum collating sequence.
7, the method for segmentation of high resolution remote sensing image based on veins clustering constrain according to claim 1, it is characterized in that, described global optimum is meant: cut apart the order that time domain merges by veins clustering pre-service influence, the zone is merged according to the direction of texture homogeneity.
8, the method for segmentation of high resolution remote sensing image based on veins clustering constrain according to claim 5, it is characterized in that, the described optimum zone that merges, be meant: in RAG figure, when the combined distance of regional A and certain adjacent area B during, then claim adjacent area B to merge the zone for the optimum in this zone less than the combined distance of regional A and other adjacent area.
9, the method for segmentation of high resolution remote sensing image based on veins clustering constrain according to claim 5 is characterized in that, described optimum combined distance is meant: in NNG figure, the optimum each other zone that merges the zone between combined distance.
10, according to claim 5 or 6 described method for segmentation of high resolution remote sensing image based on veins clustering constrain, it is characterized in that, described optimum collating sequence, be meant: in NNG figure, the optimum each other zone that merges the zone is regional right to be that local optimum in the image merges, the collating sequence that all local optimum zones are obtained ordering.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009100463483A CN101510310B (en) | 2009-02-19 | 2009-02-19 | Method for segmentation of high resolution remote sensing image based on veins clustering constrain |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009100463483A CN101510310B (en) | 2009-02-19 | 2009-02-19 | Method for segmentation of high resolution remote sensing image based on veins clustering constrain |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101510310A true CN101510310A (en) | 2009-08-19 |
| CN101510310B CN101510310B (en) | 2010-12-29 |
Family
ID=41002701
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2009100463483A Expired - Fee Related CN101510310B (en) | 2009-02-19 | 2009-02-19 | Method for segmentation of high resolution remote sensing image based on veins clustering constrain |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101510310B (en) |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102005034A (en) * | 2010-12-01 | 2011-04-06 | 南京大学 | Remote sensing image segmentation method based on region clustering |
| CN101840581B (en) * | 2010-02-05 | 2011-12-28 | 北京交通大学 | Method for extracting profile of building from satellite remote sensing image |
| CN102496151A (en) * | 2011-12-08 | 2012-06-13 | 南京大学 | Method for multi-scale segmentation of high-resolution remote sensing images |
| CN102509273A (en) * | 2011-11-21 | 2012-06-20 | 电子科技大学 | Tumor segmentation method based on homogeneous pieces and fuzzy measure of breast ultrasound image |
| CN103578110A (en) * | 2013-11-12 | 2014-02-12 | 河海大学 | Multi-band high-resolution remote sensing image segmentation method based on gray scale co-occurrence matrix |
| CN103971364A (en) * | 2014-04-04 | 2014-08-06 | 西南交通大学 | Remote sensing image variation detecting method on basis of weighted Gabor wavelet characteristics and two-stage clusters |
| CN104361589A (en) * | 2014-11-12 | 2015-02-18 | 河海大学 | High-resolution remote sensing image segmentation method based on inter-scale mapping |
| CN105513060A (en) * | 2015-11-30 | 2016-04-20 | 中国人民解放军海军航空工程学院 | Visual perception enlightening high-resolution remote-sensing image segmentation method |
| CN105719266A (en) * | 2014-12-02 | 2016-06-29 | 马鞍山森格电子科技有限公司 | Method for detecting surface defect of magnetic shoe based on texture feature clustering |
| CN106296680A (en) * | 2016-08-08 | 2017-01-04 | 长安大学 | A kind of multiple features fusion high-resolution remote sensing image dividing method based on region |
| CN109299295A (en) * | 2018-09-04 | 2019-02-01 | 南通科技职业学院 | Indigo printing fabric image database search method |
| CN109800779A (en) * | 2015-12-10 | 2019-05-24 | 河海大学 | Utilize the change detecting method of D-S evidence theory fusion FCM algorithm |
| CN109820490A (en) * | 2018-09-04 | 2019-05-31 | 宓保宏 | A kind of region symmetry method for objectively evaluating based on infrared thermal imaging technique |
| CN110135428A (en) * | 2019-04-11 | 2019-08-16 | 北京航空航天大学 | Image segmentation processing method and device |
| CN110223378A (en) * | 2019-05-28 | 2019-09-10 | 中国科学院自动化研究所 | Quadratic surface primitive extracting method, system, device based on hierarchical clustering |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1251157C (en) * | 2002-12-27 | 2006-04-12 | 中国科学院自动化研究所 | Object three-dimensional model quick obtaining method based on active vision |
| CN100378752C (en) * | 2005-06-03 | 2008-04-02 | 中国科学院自动化研究所 | Segmentation method of natural image in robustness |
| CN101211355B (en) * | 2006-12-30 | 2010-05-19 | 中国科学院计算技术研究所 | Image inquiry method based on clustering |
| CN101159009B (en) * | 2007-11-09 | 2010-04-21 | 西北工业大学 | Method for detecting bridge from remote sense image |
| CN101320467B (en) * | 2008-05-16 | 2010-06-02 | 西安电子科技大学 | Multi-dimension texture image partition method based on self-adapting window fixing and propagation |
-
2009
- 2009-02-19 CN CN2009100463483A patent/CN101510310B/en not_active Expired - Fee Related
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101840581B (en) * | 2010-02-05 | 2011-12-28 | 北京交通大学 | Method for extracting profile of building from satellite remote sensing image |
| CN102005034A (en) * | 2010-12-01 | 2011-04-06 | 南京大学 | Remote sensing image segmentation method based on region clustering |
| CN102005034B (en) * | 2010-12-01 | 2012-07-04 | 南京大学 | Remote sensing image segmentation method based on region clustering |
| CN102509273A (en) * | 2011-11-21 | 2012-06-20 | 电子科技大学 | Tumor segmentation method based on homogeneous pieces and fuzzy measure of breast ultrasound image |
| CN102509273B (en) * | 2011-11-21 | 2013-09-04 | 电子科技大学 | Tumor segmentation method based on homogeneous pieces and fuzzy measure of breast ultrasound image |
| CN102496151A (en) * | 2011-12-08 | 2012-06-13 | 南京大学 | Method for multi-scale segmentation of high-resolution remote sensing images |
| CN103578110A (en) * | 2013-11-12 | 2014-02-12 | 河海大学 | Multi-band high-resolution remote sensing image segmentation method based on gray scale co-occurrence matrix |
| CN103578110B (en) * | 2013-11-12 | 2016-06-08 | 河海大学 | Multiband high-resolution remote sensing image dividing method based on gray level co-occurrence matrixes |
| CN103971364B (en) * | 2014-04-04 | 2017-02-01 | 西南交通大学 | Remote sensing image variation detecting method on basis of weighted Gabor wavelet characteristics and two-stage clusters |
| CN103971364A (en) * | 2014-04-04 | 2014-08-06 | 西南交通大学 | Remote sensing image variation detecting method on basis of weighted Gabor wavelet characteristics and two-stage clusters |
| CN104361589A (en) * | 2014-11-12 | 2015-02-18 | 河海大学 | High-resolution remote sensing image segmentation method based on inter-scale mapping |
| CN105719266A (en) * | 2014-12-02 | 2016-06-29 | 马鞍山森格电子科技有限公司 | Method for detecting surface defect of magnetic shoe based on texture feature clustering |
| CN105719266B (en) * | 2014-12-02 | 2018-08-28 | 安徽达特智能科技有限公司 | Magnetic tile surface defect detection method based on textural characteristics cluster |
| CN105513060A (en) * | 2015-11-30 | 2016-04-20 | 中国人民解放军海军航空工程学院 | Visual perception enlightening high-resolution remote-sensing image segmentation method |
| CN109800779A (en) * | 2015-12-10 | 2019-05-24 | 河海大学 | Utilize the change detecting method of D-S evidence theory fusion FCM algorithm |
| CN109800779B (en) * | 2015-12-10 | 2020-06-30 | 河海大学 | Change detection method for fusing FCM algorithm by using D-S evidence theory |
| CN106296680A (en) * | 2016-08-08 | 2017-01-04 | 长安大学 | A kind of multiple features fusion high-resolution remote sensing image dividing method based on region |
| CN106296680B (en) * | 2016-08-08 | 2017-09-01 | 长安大学 | A kind of multiple features fusion high-resolution remote sensing image dividing method based on region |
| CN109299295A (en) * | 2018-09-04 | 2019-02-01 | 南通科技职业学院 | Indigo printing fabric image database search method |
| CN109820490A (en) * | 2018-09-04 | 2019-05-31 | 宓保宏 | A kind of region symmetry method for objectively evaluating based on infrared thermal imaging technique |
| CN109299295B (en) * | 2018-09-04 | 2021-09-28 | 南通科技职业学院 | Blue printing layout database searching method |
| CN110135428A (en) * | 2019-04-11 | 2019-08-16 | 北京航空航天大学 | Image segmentation processing method and device |
| CN110135428B (en) * | 2019-04-11 | 2021-06-04 | 北京航空航天大学 | Image segmentation processing method and device |
| CN110223378A (en) * | 2019-05-28 | 2019-09-10 | 中国科学院自动化研究所 | Quadratic surface primitive extracting method, system, device based on hierarchical clustering |
| CN110223378B (en) * | 2019-05-28 | 2021-06-22 | 中国科学院自动化研究所 | Quadric surface element extraction method, system and device based on hierarchical clustering |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101510310B (en) | 2010-12-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101510310B (en) | Method for segmentation of high resolution remote sensing image based on veins clustering constrain | |
| Weng | Remote sensing and GIS integration | |
| Cheng et al. | Accurate urban road centerline extraction from VHR imagery via multiscale segmentation and tensor voting | |
| Zhang et al. | Boundary-constrained multi-scale segmentation method for remote sensing images | |
| Dheeravath et al. | Irrigated areas of India derived using MODIS 500 m time series for the years 2001–2003 | |
| Qin et al. | Object-based land cover change detection for cross-sensor images | |
| Ma et al. | Cultivated land information extraction from high-resolution unmanned aerial vehicle imagery data | |
| Zhang et al. | Data fusion and classifier ensemble techniques for vegetation mapping in the coastal Everglades | |
| CN106846322B (en) | The SAR image segmentation method learnt based on curve wave filter and convolutional coding structure | |
| CN104881868A (en) | Method for extracting phytocoenosium spatial structure | |
| CN102810158A (en) | High-resolution remote sensing target extraction method based on multi-scale semantic model | |
| Chen et al. | Optimal segmentation of a high-resolution remote-sensing image guided by area and boundary | |
| Dong et al. | Selection of LiDAR geometric features with adaptive neighborhood size for urban land cover classification | |
| Lin et al. | Semi-automatic extraction of road networks by least squares interlaced template matching in urban areas | |
| Deliry et al. | Extracting urban impervious surfaces from Sentinel-2 and Landsat-8 satellite data for urban planning and environmental management | |
| CN104408463A (en) | High-resolution construction land graph spot identification method based on PanTex and linear characteristic | |
| CN105447452A (en) | Remote sensing sub-pixel mapping method based on spatial distribution characteristics of features | |
| CN113223042A (en) | Intelligent acquisition method and equipment for remote sensing image deep learning sample | |
| Zhang et al. | Impervious surface extraction from high-resolution satellite image using pixel-and object-based hybrid analysis | |
| Li et al. | Exploring multiple crowdsourced data to learn deep convolutional neural networks for road extraction | |
| Zhao et al. | On the topographic entity-oriented digital elevation model construction method for urban area land surface | |
| Zhou et al. | Green urban garden landscape simulation platform based on high-resolution image recognition technology and GIS | |
| Kim et al. | Generation of a DTM and building detection based on an MPF through integrating airborne lidar data and aerial images | |
| CN102360496A (en) | SAR (synthetic aperture radar) image segmentation method based on GMM (gaussian mixture model) parameter transferring and clustering | |
| Sun et al. | Synergistic use of Landsat TM and SPOT5 imagery for object-based forest classification |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20101229 Termination date: 20140219 |