CN109615637A - A kind of improved remote sensing image Hybrid Techniques - Google Patents
A kind of improved remote sensing image Hybrid Techniques Download PDFInfo
- Publication number
- CN109615637A CN109615637A CN201910038604.8A CN201910038604A CN109615637A CN 109615637 A CN109615637 A CN 109615637A CN 201910038604 A CN201910038604 A CN 201910038604A CN 109615637 A CN109615637 A CN 109615637A
- Authority
- CN
- China
- Prior art keywords
- merging
- remote sensing
- sensing image
- follows
- formula
- 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
Links
- 238000010237 hybrid technique Methods 0.000 title claims abstract description 11
- 230000011218 segmentation Effects 0.000 claims abstract description 47
- 238000001228 spectrum Methods 0.000 claims abstract description 22
- 238000004458 analytical method Methods 0.000 claims abstract description 8
- 238000012804 iterative process Methods 0.000 claims abstract description 4
- 238000000034 method Methods 0.000 claims description 37
- 238000004364 calculation method Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 230000001186 cumulative effect Effects 0.000 claims description 2
- 238000003709 image segmentation Methods 0.000 abstract description 11
- 238000000638 solvent extraction Methods 0.000 abstract description 3
- 238000005516 engineering process Methods 0.000 description 4
- 230000003595 spectral effect Effects 0.000 description 4
- 238000001514 detection method Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- VMXUWOKSQNHOCA-UKTHLTGXSA-N ranitidine Chemical compound [O-][N+](=O)\C=C(/NC)NCCSCC1=CC=C(CN(C)C)O1 VMXUWOKSQNHOCA-UKTHLTGXSA-N 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000005267 amalgamation Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000007261 regionalization Effects 0.000 description 1
- 150000003839 salts Chemical class 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of improved remote sensing image Hybrid Techniques, comprising: utilizes multi-spectrum remote sensing image data, carries out watershed transform, obtains initial segmentation result；The initial segmentation result of acquisition constructs cutting object syntople using region neighbor map RAG and closest figure NNG；The object syntople of acquisition, considers heterogeneous between homogeney and object in object, calculates the merging cost between each adjacent object；Merging threshold is set, if the minimum threshold value for merging cost and being less than definition, carries out object merging；RAG and NNG is updated, new object syntople is obtained, when merging cost more than or equal to threshold value, stops iterative process, and export final segmentation result.Compared with existing Remote Sensing Image Segmentation, the present invention utilizes image edge information and spatial information, over-segmentation/less divided the problem for reducing previous partitioning algorithm effectively realizes the good segmentation of multi-spectrum remote sensing image as a result, providing good basis for further remote sensing images analysis.
Description
Technical field
The invention belongs to the technical field of Remote Sensing Image Segmentation, in particular to a kind of improved remote sensing image mixes segmentation side
Method.
Background technique
Along with the fast development of multiresolution remote sensing satellite technology, application of the Object-oriented Technique in remote sensing analysis is got over
Come more universal, object-oriented, which refers to, utilizes spectrum, texture and contextual information and geographic object feature to carry out image classification.With biography
The pixels approach of system compares, and object-oriented is lower to the susceptibility of spectral variables in geographic object, therefore can reduce " spiced salt "
Noise.Object-oriented Technique is related to Object identifying, the processes such as information extraction and image classification, wherein image object be towards
The key foundation of object analysis technology.Therefore, image segmentation is carried out to obtain accurate image object, is object-oriented analysis skill
The important prerequisite of art.
When classifying to remote sensing image, image segmentation is typically considered a complicated job, currently, remote sensing shadow
As dividing method has obtained definitely developing.The method of image segmentation is generally divided into two kinds: segmentation based on edge and being based on
The segmentation in region.Dividing method based on edge thinks that gray value is discontinuous, therefore search graph at different boundary region
The discontinuous position of gray value is as in determine edge.And the dividing method based on region considers the similitude and phase between pixel
Adjacent relationship, to ensure that every an object of Image Segmentation meets certain similarity measurement.But above-mentioned dividing method often generated
Segmentation or less divided, are not able to satisfy the needs of remote sensing application.
Now, it divides and is considered as a kind of new trend to solve the above problems with the method that merging combines, that is, first with
Initial segmentation result is obtained based on the dividing method at edge, and the method based on region is then utilized to merge similar segmentation pair
As.This method considers the spatial information between the boundary information of remote sensing image internal object and adjacent geographic object simultaneously.
Region merging technique is a kind of effective method for Remote Sensing Image Segmentation, because it can use certain homogeneys or heterogeneity
Measurement.Merging sequence and merging criterion are the critical issue during region merging technique, recently, the Image Segmentation based on merging sequence
Algorithm is published successively.For example, large-scale heterogeneous agricultural regionalization is divided into tool by Canovas-Garcia and Alonso-Sarria
There is the plot of different Land cover types, to select best scale parameter.Yang et al. proposes a kind of novel mixing point
Segmentation method, this method carry out region merging technique using local spectral modeling degree threshold value, and this method is in target detection, object-based classification
There is very big application potential with detection etc. is changed.Although the method based on merging sequence (MO) shows in image segmentation
Well, but these methods all rely on merging criterion (MC) to determine amalgamation result.Therefore, constructing merging criterion MC appropriate is
The key of region merging technique.
Summary of the invention
The technical problem to be solved in the present invention are as follows: overcome the deficiencies in the prior art, it is frequent in order to reduce image division method
The problem of over-segmentation of appearance or less divided, provides a kind of improved remote sensing image Hybrid Techniques, has effectively achieved more
The good segmentation of spectral remote sensing image is as a result, provide good basis, realization remote sensing image for further remote sensing images analysis
Geographic object target accurately identifies.
The present invention solves the technical solution that above-mentioned technical problem uses are as follows: a kind of improved remote sensing image mixing segmentation side
Method, steps are as follows:
Step 1: carrying out watershed transform using multi-spectrum remote sensing image data, initial segmentation result is obtained；
Step 2: utilizing region neighbor map RAG and closest figure NNG structure based on the initial segmentation result that step 1 obtains
Build cutting object syntople；
Step 3: based on the object syntople that step 2 obtains, according to heterogeneous, meter between homogeney and object in object
Calculate the merging cost between each adjacent object；
Step 4: setting merging threshold, if the minimum merging threshold for merging cost and being less than setting, carries out object conjunction
And；
Step 5: updating RAG and NNG, new object syntople is obtained, repeats step 3 and step 4, when merging generation
When valence is greater than or equal to the merging threshold of setting, stop iterative process, and export final segmentation result.
In the step 1, the process for obtaining initial segmentation result is:
(1) according to watershed transform theory, multi-spectrum remote sensing image gradient value is calculated using sobel operator, and to mostly light
It composes remote sensing gradient image and carries out wave band weighted average, calculation formula are as follows:
In formula, p indicates a pixel in a width image, and k indicates wave band, fk(p) the DN value of pixel p in K-band is indicated,
N represents the wave band number of remote sensing image, and sobel operator representation isIt is a 3*3 matrix；
(2) watershed transform is carried out to based on the average weighted gradient image of wave band, to obtain initial segmentation result.
In the step 2, the process of region neighbor map RAG and closest figure NNG building cutting object syntople is utilized
It is as follows:
(1) initial segmentation result obtained based on step 1, using the syntople of RAG building cutting object, and with number
Group mode (N*2) stores adjacent object pair, and N represents total adjacent object number.
(2) using the merging cost of corresponding adjacent object pair in NNG storage RAG, storage mode is array mode (N*
1)。
In the step 3, the merging cost between each adjacent object is calculated, the process for calculating merging cost is:
(1) the object syntople obtained based on step 2 calculates the spectrum intervals of contiguous object using Euclidean distance ED,
To measure the heterogeneity of the spectrum between contiguous object, Euclidean distance ED is indicated are as follows:
EDi,j=| | ui-uj||2 (2)
In formula, i and j represent a pair of of adjacent segmentation object in the initial segmentation result obtained based on step 1, uiAnd ujIt is
The spectrum average of adjacent object i and j, ‖ ui-uj‖ is the Euclidean distance of adjacent object i and j；
(2) area and the public boundary weighting for calculating contiguous object are heterogeneousIt indicates are as follows:
In formula, AiAnd AjIt is the area of object i and object j respectively, L is the public boundary of i and j, and area is small or public boundary
Big adjacent object has high merging priority；
(3) in computing object the standard deviation (STDV) of DN value to reflect the spectrum homogeney of each object, wave band weighting
Homogeney HSIt indicates are as follows:
In formula, S is a cutting object in initial segmentation result, and p is a pixel in object S, DNi(p) it indicates
The DN value of pixel p in wave band i；
(4) it is weighted using the area of each cutting object, calculates the overall homogeneity of cutting objectIt indicates are as follows:
(5) the individual homogeney of cutting object and the ratio of overall homogeneity are calculated, to obtain opposite homogeney RH, relatively
Homogeney indicates are as follows:
In formula, RH value is bigger, indicates that the homogeney in object S is higher；
(6) formula (3) and formula (6) are combined, cost calculating is merged, the merging cost t of each adjacent object is indicated are as follows:
In the step 4, the merging cost based on each adjacent object that step 3 obtains, setting merging threshold T is carried out
Merge, threshold value T is obtained by accumulated probability analysis, is indicated are as follows:
F(xa)=P (x≤T)=α, 0 < α < 1 (8)
In formula, x is the merging cost value of adjacent object, and P is the cumulative probability as x≤T, and each α corresponds to unique
T, and when α increases, T increases.
The advantages of the present invention over the prior art are that:
(1) through the invention the step of, realizes remote sensing image mixing segmentation, effectively improves traditional Image Segmentation and goes out
Existing over-segmentation and less divided problem, provides premise guarantee for Object-oriented Technology.
(2) proposed by the present invention that remote sensing shadow is significantly reduced based on object homogeney and heterogeneous region merging method
As the error in segmentation research, the marginal information and spatial information of image is adequately utilized, reduces previous partitioning algorithm
Over-segmentation/less divided problem improves image precision, realizes the combination of new technology and innovation research.
(3) a kind of improved remote sensing image mixing partitioning algorithm proposed by the present invention has fully considered the edge letter of image
Breath and spatial information using the local autocorrelation performance of image can divide the geographic object of different sizes, illustrate closing
And it is heterogeneous between homogeney and object in consideration object in the process, be conducive to improve segmentation precision, while there is strong applicability, meter
Simple feature can be realized the segmentation of most of land types.
Detailed description of the invention
Fig. 1 is the overall procedure schematic diagram that the present invention carries out remote sensing image mixing segmentation；
Fig. 2 is the flow diagram that the present invention merges that cost calculates；
Fig. 3 is the segmentation result figure that factory is directed in present example.
Specific embodiment
It is tested below in conjunction with One District In Beijing domain Remote Sensing Image Segmentation, to a kind of improved remote sensing proposed by the present invention
Image Hybrid Techniques are described in detail.
The present invention is embodied in example, and the high score No.1 remote sensing image spatial resolution used is 2 meter；It uses
The visual fusion of NNDiffuse Pan Sharpening function and high score No.1 8m resolution ratio in ENVI5.2 software, to increase
Strong spectral information；Fused image includes blue, green, red and four wave bands of near-infrared.
As shown in Figure 1, a kind of improved remote sensing image Hybrid Techniques of the present invention, realize that steps are as follows:
Step (A) utilizes multi-spectrum remote sensing image data, progress watershed transform, acquisition initial segmentation result.The step
It is mainly realized by watershed transform, embodiment are as follows:
A1. multi-spectrum remote sensing image gradient value is calculated using sobel operator, and wave is carried out to multispectral remote sensing gradient image
Section weighted average, calculation formula are as follows:
In formula, p indicates a pixel in a width image, fk(p) the DN value of pixel p in K-band is indicated, n represents remote sensing
The wave band number of image, sobel operator representation areIt is a 3*3 matrix；
A2. watershed transform is carried out to based on the average weighted gradient image of wave band, to obtain initial segmentation result.
Step (B), the initial segmentation result obtained based on step (A), utilize region neighbor map RAG and closest figure NNG
Construct cutting object syntople；
Step (C), the object syntople obtained based on step (B), consider it is heterogeneous between homogeney and object in object,
Calculate the merging cost between each adjacent object.The embodiment that the step calculates merging cost is as shown in Figure 2:
C1. Euclidean distance ED is calculated as spectrum intervals, to measure the heterogeneity of the spectrum between contiguous object, Euclidean distance
It may be expressed as:
EDi,j=| | ui-uj||2 (10)
I and j represents a pair of of adjacent segmentation object in the initial segmentation result obtained based on step (A), uiAnd ujIt is adjacent
The spectrum average of object i and j, ‖ ui-uj‖ is the Euclidean distance of adjacent object i and j；
C2. area and the public boundary weighting for calculating contiguous object are heterogeneousIt indicates are as follows:
In formula, AiAnd AjIt is the area of object i and object j respectively, L is the public boundary of i and j, and area is small or public boundary
Big adjacent object has high merging priority；
C3. in computing object the standard deviation (STDV) of DN value to reflect the spectrum homogeney of each object, wave band weighting
Homogeney HSIt can indicate are as follows:
In formula, S is a cutting object in initial segmentation result, and p is a pixel in object S, DNi(p) it indicates
The DN value of pixel p in wave band i；
C4. it is weighted using the area of each cutting object, calculates the overall homogeneity of cutting object, indicated are as follows:
C5. the individual homogeney of cutting object and the ratio of overall homogeneity are calculated, to obtain opposite homogeney RH, relatively
Homogeney indicates are as follows:
In formula, RH value is bigger, indicates that the homogeney in object S is higher；
C6. convolution (11) and formula (14), merge cost calculating, and the merging cost t of each adjacent object is indicated are as follows:
Step (D), setting merging threshold T, if the minimum threshold value for merging cost and being less than definition, carries out object merging,
Threshold value T is obtained by accumulated probability analysis, is indicated are as follows:
F(xa)=P (x≤T)=α, 0 < α < 1 (16)
In the embodiment of the present invention, α is set as 0.6.
Step (E) updates RAG and NNG, obtains new object syntople, repeats step (C) and step (D), works as merging
When cost is greater than or equal to threshold value, stop iterative process, and export final segmentation result.
In embodiments of the present invention, (a) is original multi-spectrum remote sensing image to be split in Fig. 3, is merged using the present invention
Algorithm, corresponding segmentation result are shown in (b) in Fig. 3, and it is that two kinds of merging algorithms of OH and FLSA exist that (c) and (d) is corresponding in Fig. 3
α is set as 0.6 segmentation result.Compared with (b) in Fig. 3, there is different degrees of mistake in atural object in (c) and (d) in Fig. 3
Divide phenomenon.It can be seen that inventive algorithm realizes the good segmentation of remote sensing image atural object.
Non-elaborated part of the present invention belongs to techniques well known.
The above, part specific embodiment only of the present invention, but scope of protection of the present invention is not limited thereto, appoints
In the technical scope disclosed by the present invention, any changes or substitutions that can be easily thought of, should all cover by what those skilled in the art
Within protection scope of the present invention.
Claims (5)
1. a kind of improved remote sensing image Hybrid Techniques, which is characterized in that this method realizes that steps are as follows:
Step 1: carrying out watershed transform using multi-spectrum remote sensing image data, initial segmentation result is obtained；
Step 2: utilizing region neighbor map RAG and closest figure NNG building point based on the initial segmentation result that step 1 obtains
Cut object syntople；
Step 3: based on the object syntople that step 2 obtains, according to heterogeneity, calculating are every between homogeney and object in object
Merging cost between a adjacent object；
Step 4: setting merging threshold, if the minimum merging threshold for merging cost and being less than setting, carries out object merging；
Step 5: updating RAG and NNG, new object syntople is obtained, repeats step 3 and step 4, cost is big when merging
In or equal to setting merging threshold when, stop iterative process, and export final segmentation result.
2. a kind of improved remote sensing image Hybrid Techniques according to claim 1, it is characterised in that: the step 1
In, the process for obtaining initial segmentation result is:
(1) according to watershed transform theory, multi-spectrum remote sensing image gradient value is calculated using sobel operator, and to multispectral distant
Feel gradient image and carry out wave band weighted average, calculation formula are as follows:
In formula, p indicates a pixel in a width image, and k indicates wave band, fk(p) the DN value of pixel p in K-band is indicated, n is represented
The wave band number of remote sensing image, sobel operator representation areIt is a 3*3 matrix；
(2) watershed transform is carried out to based on the average weighted gradient image of wave band, to obtain initial segmentation result.
3. a kind of improved remote sensing image Hybrid Techniques according to claim 1, it is characterised in that: state step 2
In, the process using region neighbor map RAG and closest figure NNG building cutting object syntople is as follows:
(1) initial segmentation result obtained based on step 1, using the syntople of RAG building cutting object, and with array side
Formula (N*2) stores adjacent object pair, and N represents total adjacent object number；
(2) using the merging cost of corresponding adjacent object pair in NNG storage RAG, storage mode is array mode (N*1).
4. a kind of improved remote sensing image Hybrid Techniques according to claim 1, it is characterised in that: the step 3
In, the merging cost between each adjacent object is calculated, the process for calculating merging cost is:
(1) the object syntople obtained based on step 2 calculates the spectrum intervals of contiguous object using Euclidean distance ED, to survey
The spectrum measured between contiguous object is heterogeneous, and Euclidean distance ED is indicated are as follows:
EDi,j=| | ui-uj||2 (2)
In formula, i and j represent a pair of of adjacent segmentation object in the initial segmentation result obtained based on step 1, uiAnd ujIt is adjacent
The spectrum average of object i and j, ‖ ui-uj‖ is the Euclidean distance of adjacent object i and j；
(2) area and the public boundary weighting for calculating contiguous object are heterogeneousIt indicates are as follows:
In formula, AiAnd AjIt is the area of object i and object j respectively, L is the public boundary of i and j, and area is small or public boundary is big
Adjacent object has high merging priority；
(3) for the standard deviation (STDV) of DN value to reflect the spectrum homogeney of each object, wave band weights homogeneity in computing object
Property HSIt indicates are as follows:
In formula, S is a cutting object in initial segmentation result, and p is a pixel in object S, DNi(p) wave band i is indicated
In pixel p DN value；
(4) it is weighted using the area of each cutting object, calculates the overall homogeneity of cutting objectIt indicates are as follows:
(5) the individual homogeney of cutting object and the ratio of overall homogeneity are calculated, to obtain opposite homogeney RH, opposite homogeneity
Property indicate are as follows:
In formula, RH value is bigger, indicates that the homogeney in object S is higher；
(6) formula (3) and formula (6) are combined, cost calculating is merged, the merging cost t of each adjacent object is indicated are as follows:
5. a kind of improved remote sensing image Hybrid Techniques according to claim 1, it is characterised in that: the step 4
In, the merging cost based on each adjacent object that step 3 obtains, setting merging threshold T is merged, and threshold value T passes through accumulative
Probability analysis obtains, indicates are as follows:
F(xa)=P (x≤T)=α, 0 < α < 1 (8)
In formula, x is the merging cost value of adjacent object, and P is the cumulative probability as x≤T, and each α corresponds to unique T, and
And T increases when α increases.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910038604.8A CN109615637A (en) | 2019-01-16 | 2019-01-16 | A kind of improved remote sensing image Hybrid Techniques |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910038604.8A CN109615637A (en) | 2019-01-16 | 2019-01-16 | A kind of improved remote sensing image Hybrid Techniques |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109615637A true CN109615637A (en) | 2019-04-12 |
Family
ID=66019772
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910038604.8A Pending CN109615637A (en) | 2019-01-16 | 2019-01-16 | A kind of improved remote sensing image Hybrid Techniques |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109615637A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110910397A (en) * | 2019-10-18 | 2020-03-24 | 中国人民解放军陆军工程大学 | Remote sensing image segmentation method |
| CN111868783A (en) * | 2019-02-14 | 2020-10-30 | 中国水利水电科学研究院 | Region merging image segmentation algorithm based on boundary extraction |
| CN112070745A (en) * | 2020-09-08 | 2020-12-11 | 中国科学院空天信息创新研究院 | Rapid and effective remote sensing image segmentation non-supervision evaluation method |
| CN115239746A (en) * | 2022-09-23 | 2022-10-25 | 成都国星宇航科技股份有限公司 | Object-oriented remote sensing image segmentation method, device, equipment and medium |
| CN116681711A (en) * | 2023-04-25 | 2023-09-01 | 中国科学院地理科学与资源研究所 | Multi-scale segmentation method for high-resolution remote sensing image under partition guidance |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101425182A (en) * | 2008-11-28 | 2009-05-06 | 华中科技大学 | Image object segmentation method |
| CN104112007A (en) * | 2014-07-16 | 2014-10-22 | 深圳大学 | Data storage, organization and retrieval methods of image gradation segmentation result |
| CN108090913A (en) * | 2017-12-12 | 2018-05-29 | 河南大学 | A kind of image, semantic dividing method based on object level Gauss-Markov random fields |
-
2019
- 2019-01-16 CN CN201910038604.8A patent/CN109615637A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101425182A (en) * | 2008-11-28 | 2009-05-06 | 华中科技大学 | Image object segmentation method |
| CN104112007A (en) * | 2014-07-16 | 2014-10-22 | 深圳大学 | Data storage, organization and retrieval methods of image gradation segmentation result |
| CN108090913A (en) * | 2017-12-12 | 2018-05-29 | 河南大学 | A kind of image, semantic dividing method based on object level Gauss-Markov random fields |
Non-Patent Citations (1)
| Title |
|---|
| YONGJI WANG等: "Region Merging Considering Within- and Between-Segment Heterogeneity: An Improved Hybrid Remote-Sensing Image Segmentation Method", 《REMOTE SENSING》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111868783A (en) * | 2019-02-14 | 2020-10-30 | 中国水利水电科学研究院 | Region merging image segmentation algorithm based on boundary extraction |
| CN111868783B (en) * | 2019-02-14 | 2021-03-23 | 中国水利水电科学研究院 | Region merging image segmentation algorithm based on boundary extraction |
| CN110910397A (en) * | 2019-10-18 | 2020-03-24 | 中国人民解放军陆军工程大学 | Remote sensing image segmentation method |
| CN110910397B (en) * | 2019-10-18 | 2022-10-11 | 中国人民解放军陆军工程大学 | Remote sensing image segmentation method |
| CN112070745A (en) * | 2020-09-08 | 2020-12-11 | 中国科学院空天信息创新研究院 | Rapid and effective remote sensing image segmentation non-supervision evaluation method |
| CN112070745B (en) * | 2020-09-08 | 2022-11-22 | 中国科学院空天信息创新研究院 | Rapid and effective remote sensing image segmentation non-supervision evaluation method |
| CN115239746A (en) * | 2022-09-23 | 2022-10-25 | 成都国星宇航科技股份有限公司 | Object-oriented remote sensing image segmentation method, device, equipment and medium |
| CN116681711A (en) * | 2023-04-25 | 2023-09-01 | 中国科学院地理科学与资源研究所 | Multi-scale segmentation method for high-resolution remote sensing image under partition guidance |
| CN116681711B (en) * | 2023-04-25 | 2024-01-30 | 中国科学院地理科学与资源研究所 | Multi-scale segmentation method for high-resolution remote sensing image under partition guidance |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109615637A (en) | A kind of improved remote sensing image Hybrid Techniques | |
| CN105139015B (en) | A kind of remote sensing images Clean water withdraw method | |
| CN107067405B (en) | Remote sensing image segmentation method based on scale optimization | |
| CN104200461B (en) | The remote sensing image registration method of block and sift features is selected based on mutual information image | |
| CN107330875B (en) | Water body surrounding environment change detection method based on forward and reverse heterogeneity of remote sensing image | |
| CN109389062B (en) | Method for extracting lake water and land dividing line by using high-resolution satellite-borne SAR image | |
| CN103632363A (en) | Object-level high-resolution remote sensing image change detection method based on multi-scale fusion | |
| CN103971115A (en) | Automatic extraction method for newly-increased construction land image spots in high-resolution remote sensing images based on NDVI and PanTex index | |
| CN106339674A (en) | Hyperspectral image classification method based on edge preservation and graph cut model | |
| CN109859219B (en) | High-resolution remote sensing image segmentation method combining phase and spectrum | |
| CN103279951A (en) | Object-oriented remote sensing image building and shade extraction method of remote sensing image building | |
| CN102903109B (en) | A kind of optical image and SAR image integration segmentation method for registering | |
| CN109363699A (en) | A kind of method and device of breast image lesion identification | |
| Zhu et al. | Robust registration of aerial images and LiDAR data using spatial constraints and Gabor structural features | |
| US11238307B1 (en) | System for performing change detection within a 3D geospatial model based upon semantic change detection using deep learning and related methods | |
| US11587249B2 (en) | Artificial intelligence (AI) system and methods for generating estimated height maps from electro-optic imagery | |
| CN109363697A (en) | A kind of method and device of breast image lesion identification | |
| US20220091259A1 (en) | System using a priori terrain height data for interferometric synthetic aperture radar (ifsar) phase disambiguation and related methods | |
| CN103839267A (en) | Building extracting method based on morphological building indexes | |
| CN114005042A (en) | Remote sensing image urban building extraction method based on shadow compensation and U-net | |
| CN110070545A (en) | A kind of method that textural characteristics density in cities and towns automatically extracts cities and towns built-up areas | |
| Fan et al. | Region-merging method with texture pattern attention for SAR image segmentation | |
| CN106650749B (en) | The drawing method of right angle building in a kind of high-resolution optical image | |
| Fang et al. | Supervoxel-based brain tumor segmentation with multimodal MRI images | |
| CN112734683B (en) | Multi-scale SAR and infrared image fusion method based on target enhancement |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190412 |