CN112241956A - PolSAR image ridge line extraction method based on region growing method and variation function - Google Patents
PolSAR image ridge line extraction method based on region growing method and variation function Download PDFInfo
- Publication number
- CN112241956A CN112241956A CN202011206437.2A CN202011206437A CN112241956A CN 112241956 A CN112241956 A CN 112241956A CN 202011206437 A CN202011206437 A CN 202011206437A CN 112241956 A CN112241956 A CN 112241956A
- Authority
- CN
- China
- Prior art keywords
- image
- ridge line
- building
- texture
- features
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/40—Analysis of texture
- G06T7/41—Analysis of texture based on statistical description of texture
-
- 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
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- 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
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20092—Interactive image processing based on input by user
- G06T2207/20104—Interactive definition of region of interest [ROI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
Abstract
The invention discloses a PolSAR image ridge line extraction method based on a region growing method and a variation function, which comprises the steps of firstly, initially segmenting a PolSAR image by using a region growing method for automatically selecting seed points to obtain a mixed ground object of a ridge line and a building; secondly, calculating an omnidirectional variation function of the image to obtain variation texture features, calculating texture entropy features of the image, superposing the variation texture features and the texture entropy features to obtain combined texture features, performing cluster analysis on the combined texture features by adopting a fuzzy C mean clustering method, and extracting a building region; and finally, combining the ridge line and the mixed ground object of the building with the extracted building area to obtain a ridge line identification result. The method for automatically selecting the seed points has strong anti-interference capability, obtains higher ridge line identification precision by utilizing the textural features of the PolSAR image, and provides a new idea for ridge line identification of the PolSAR image.
Description
Technical Field
The invention belongs to the technical field of geographic information, and particularly relates to a PolSAR image ridge line extraction method based on a region growing method and a variation function.
Background
The mountain disaster triggered by the earthquake usually starts near a ridge line, and the accurate acquisition of the distribution information of the ridge line is beneficial to understanding the mountain disaster conditions triggered by the earthquake, such as road damage, mountain collapse, landslide and the like, so that assistance is provided for rescue workers and material transportation. After an earthquake occurs, under the influence of overcast and rainy weather, the function of an optical remote sensing image in disaster relief decision is greatly limited, and radar remote sensing is widely used in earthquake research due to the advantages of strong penetrating power, independence on solar illumination and the like. In a Synthetic Aperture Radar (SAR) image, the features of a ridge and a building are very similar, and the existence of the building can bring great influence to the identification of the ridge, so that the ridge line identification difficulty is high. In the existing ridge line extraction method, the anti-interference capability of the region growing method of manually selecting seed points is poor when the image is initially segmented, and the influence on the initial segmentation effect of the image is large; the ridge line extracted by the traditional threshold method is low in precision, and the result accuracy is poor.
Disclosure of Invention
Aiming at the defects pointed out in the background technology, the invention provides a PolSAR image ridge line extraction method based on a region growing method and a variation function, and aims to solve the problems in the prior art in the background technology.
In order to achieve the purpose, the invention adopts the technical scheme that:
a PolSAR image ridge line extraction method based on a region growing method and a variation function comprises the following steps:
(1) initial segmentation of PolSAR image by using region growing method for automatically selecting seed points
Firstly, selecting initial seed points by combining boundary information of an image, designating one seed point as a starting point of growth, merging pixels and seed points of four fields or eight fields of the initial seed points, merging all pixels with similar properties and continuing to grow outwards until pixels which do not meet conditions are added, and obtaining a mixed ground object of a ridge line and a building;
(2) calculating an omnidirectional variation function of an image to obtain variation texture features, simultaneously calculating texture entropy features of the image, superposing the variation texture features and the texture entropy features to obtain combined texture features, performing cluster analysis on the combined texture features by adopting a fuzzy C mean value clustering method, setting two clustering centers in a clustering process, respectively corresponding to two types of ground objects, namely buildings and non-buildings, and extracting building areas after completing cluster segmentation;
(3) and (3) combining the ridge line and the mixed ground feature of the building obtained in the step (1) with the building area extracted in the step (2) to obtain a ridge line identification result graph.
Preferably, in the step (1), the automatic selection method of the initial seed point includes: traversing the whole image pixel by using a window with a certain size, calculating the pixel value mean value of other coordinate points except the central point in each window, calculating the difference value of the mean value and the pixel value of the central point, and selecting the coordinate points meeting the following three conditions as initial seed points at the same time, wherein the three conditions are as follows:
a. the calculation result of the difference value of the initial seed points is as close as possible to the median of an interval formed by 0 and the threshold value of the region growing method;
b. the initial seed point falls within a ridge line region in the image;
c. the distance between the initial seed point and the central point of the map sheet is the nearest.
Preferably, in step (2), the method for obtaining the omnidirectional variation function of the image includes: taking the width of the window as W, and setting the coordinate of each pixel in the image as (i, j), wherein the coordinate of the center point of the window is (i)0,j0) Taking the mean value of the calculation results of the variation functions in four directions of 0 °, 45 °, 90 ° and 135 ° as the omnidirectional variation function of the image, the calculation formula is as follows:
compared with the defects and shortcomings of the prior art, the invention has the following beneficial effects:
in the PolSAR image ridge line extraction method based on the region growing method and the variation function, the region growing method of automatically selecting the seed points is adopted, and compared with the method of manually selecting the seed points, the method has stronger anti-interference capability; after the image is initially segmented, variation texture features and texture entropy features of the image are calculated, the variation texture features and the texture entropy features are combined to obtain combined texture features, then a building area is extracted by adopting a fuzzy C-means clustering algorithm and is combined with an initial segmentation result, and ridge line identification is achieved. The method can distinguish the building area from the ridge, can simultaneously identify the continuous large ridge and the discontinuous small ridge, has high identification precision, greatly improves the ridge line identification precision compared with the precision of the traditional threshold segmentation method, and provides a new idea for identifying the ridge line in the PolSAR image.
Drawings
Fig. 1 is a framework flowchart of a method for extracting a polar sar image ridge line based on a region growing method and a variogram according to an embodiment of the present invention.
Fig. 2 is a mixed map of a ridge line and a building obtained by the area growth according to an embodiment of the present invention.
Fig. 3 is a calculation result (a) of the experimental variation function and an image (b) corresponding to the texture entropy feature according to the embodiment of the present invention.
Fig. 4 is a diagram of a result of identifying a building and a ridge line according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The invention provides a PolSAR image ridge line extraction method based on a region growing method and a variation function, a frame flow chart refers to FIG. 1, and the method comprises the following steps:
1. initial segmentation of PolSAR image by using region growing method for automatically selecting seed points
Selecting an initial seed point by combining boundary information of the PolSAR image, firstly, designating a seed point as a growth starting point, wherein the result of region growth is often greatly different from the actual result or the expected result due to an improper seed point, then merging pixels and seed points in four fields or eight fields of the initial seed point, merging all pixels with similar properties and continuing to grow to the outside until no pixel meeting the condition is added, and obtaining a mixed ground object of a ridge line and a building.
When the initial seed points are selected, certain conditions need to be met, namely the characteristics of the region of interest can be represented, and the region of interest in the invention refers to a ridge line region. Setting the size of any image I as M multiplied by N, wherein M and N respectively represent the length and width of the image, the coordinates of any pixel point in the image are (I, j), the pixel value of a corresponding point is I (I, j), the threshold value of a region growing method is T epsilon R,
represents the mean value of the corresponding pixel values of the eight pixel points around the point (i, j)Calculating the whole image
Difference from its central pixel I (I, j)
As shown in the formula (1),
where i ∈ (1, …, N), j ∈ (1, …, M). If the pixel corresponding to the point (i, j) is in the region of interest in the image, the image segmentation result obtained by performing region growing by using the point as the initial seed point will depend on the distance between the point and the center of the image frame and
the value changes. If it is
The closer the value is to [0-T]The closer the distance between the initial seed point and the center of the image is, the more sufficient the growth process is, the higher the efficiency is, and the better the segmentation effect is.
2. Obtaining texture information of building area by combining variation texture and texture entropy characteristics
The variation function is a method for describing spatial variation, which considers the randomness of the regionalized variable and the spatial structure characteristics of the data, and can quantitatively describe the spatial correlation of the regionalized variable. Defining the variance of the regionalized variable g (x) at points x and x + h as half of the variance of the variable g (x) in the h direction, i.e.:
if the mathematical expectation of the regionalized variable G (x) is that E [ G (x) ] exists and is equal to the constant k, and the covariance at any two points x and x + h is dependent only on h, and not on x itself, the second order stationary assumption is satisfied. Since a texture image is composed of texture units that are repeated to some extent, and thus, in accordance with this assumption, the variation function σ (h) and the covariance function cov (h) are respectively:
Cov(h)＝E[G(x)G(x+h)]-{E[G(x)]}2 (4)
this can be deduced:
σ(h)＝Cov(0)-Cov(h) (5)
h is called as the variation range of sigma (h), which reflects the influence range of regional variables, the variation function in the practical process can not be directly obtained, and the variation function is usually estimated by using an experimental variation function0,j0) Taking the mean value of the calculation results of the experimental variation functions in four directions of 0 °, 45 °, 90 ° and 135 ° as an omnidirectional variation function, and the calculation formula is as follows:
the entropy characteristic quantity with low information redundancy is selected, the texture entropy characteristic of the image is calculated, the ridge has significant characteristics in the texture entropy characteristic image, but is usually incomplete, and more fine and broken ridges are not recognized, so that a complete ridge line cannot be obtained only through the texture entropy characteristic. Therefore, the texture entropy feature and the variation texture feature need to be superimposed to obtain the combined texture feature.
3. Building region extraction by fuzzy C-means clustering algorithm
Fuzzy C-means (FCM) clustering algorithm fusionCompared with the hard clustering of K-means, the FCM provides a more flexible clustering result. The clustering loss function of FCM is shown in equation (7). The algorithm is to use t vectors xi(i-1, 2, …, t) is divided into C fuzzy clusters and the corresponding cluster center of each cluster is found, so as to minimize the cluster loss function.
Where K represents the number of samples, C represents the number of object classes, CjDenotes the center of the j-th cluster, xiDenotes the ith sample, vijAnd (3) representing the membership degree of the sample i belonging to the jth class, wherein the fuzzy C-means clustering method is a process of continuously and iteratively calculating the membership degree and the cluster center until the membership degree and the cluster center meet the condition of the formula (10).
Where l is the number of iteration steps and epsilon is the error threshold. When iteration is carried out to a certain degree, and the membership degree does not change too much, the local or global optimal state is considered to be reached.
And performing cluster analysis on the combined texture features by adopting a fuzzy C-means clustering method, setting two clustering centers in the clustering process, respectively corresponding to two types of ground objects, namely buildings and non-buildings, and extracting a building area after completing cluster segmentation.
4. Identification of ridge lines
And combining the obtained ridge line and mixed ground features of the building with the extracted building area to obtain a ridge line identification result graph.
Example 1
1. Seed point selection and image segmentation
The selection of the seed point position is closely related to the segmentation result, when the selected seed point position is better, the segmentation time is shorter, and the segmentation result is better.
(1) Traversing the Span image of the experimental area pixel by using windows with the size of 3 multiplied by 3, calculating the mean value of the pixel values of the rest points except the central point in each window, calculating the difference value between the pixel values of the rest points and the pixel value of the central point, and counting the calculation results of all the difference values to prepare for subsequent analysis.
(2) In order to obtain a better image segmentation result, a proper threshold value T needs to be set, and in the region growing process, the threshold value is used as a judgment basis for judging whether a next pixel point is added into the target region. Considering the factors such as the accuracy and robustness of threshold setting, 100 ridge samples and non-ridge samples are selected from the Span image of the research area, the gray levels of the non-ridge samples are mainly concentrated below 179, and the ridge samples are opposite to the non-ridge samples and are mostly above 179, so that the initial seed points are selected in the area larger than 179. According to experimental experience, the threshold value of the region growing algorithm is set to 76 in the embodiment. In addition, the calculation result of the difference value needs to be screened to select the optimal initial seed point for region growth, and if a certain point in the image simultaneously meets the following three conditions: (1) the difference calculation result of the point is as close as possible to the median of the interval [0,76 ]; (2) the point falls in a ridge line area in the image, namely an area with Span image gray scale larger than 179; (3) and the distance between the point and the central point of the map sheet is the shortest, so that the point is the optimal initial seed point. When the point is used as a seed point for regional growth, the growth process is sufficient, the time consumption is short, and the precision is high. Based on this requirement, a point with coordinate (651,443) on the image is finally selected as an initial seed point for region growing to segment the image, and the result shown in fig. 2 is obtained. In the figure, the white part is a mixed region of the ridge and the building, and the black part is other types of ground objects, so that the ridge and the building are difficult to distinguish by only using the region growing method.
(3) In the Span image, the gray level change of a building area is severe, the gray level change of a non-building area is relatively small, the two areas have large difference in non-similarity, and the texture information of the building is calculated by using the characteristic and adopting a variation function. Considering that the gray level change condition of a building area is relatively complex, the invention calculates the variation function in four directions of 0 degrees, 45 degrees, 90 degrees and 135 degrees. In order to keep the texture information of the building as much as possible and simultaneously maximize the distinction degree between the building and the non-building, different variable range values are firstly set for variation calculation so as to select the optimal calculation result, and experiments show that the calculation result is optimal when the variable range value is 3. At this time, the size of the corresponding calculation window is 15 × 15, the window is used for performing average variation calculation in four directions, the obtained result has the maximum discrimination between the building and the non-building, and as shown in fig. 3(a), the building area has certain texture characteristics, the correlation between the regionalized variables is large, the gray level change is severe, and the gray level change of most non-building areas is moderate, so that the characteristic value of the variation function of the building area is large relative to the non-building area, and the characteristic of high brightness is presented. The Alos-2 satellite data with medium and low resolution is adopted, the proportion of the building in the whole research area is small, meanwhile, the local change is obvious at the boundary of the building and the ridge or the boundary of the ridge and other types of ground objects, the correlation of regional variables is small, and the calculated characteristic value of the variation function of the ridge area is large due to the reasons, so that the individual ridge area is misjudged as the building. In order to reduce the influence of the mistaken division on the subsequent FCM clustering result, the texture entropy characteristics based on the gray level co-occurrence matrix are superposed with the variation function texture extraction result. When texture entropy features are extracted, the size of a calculation window is 7 multiplied by 7, the calculation directions are four directions of 0 degrees, 45 degrees, 90 degrees and 135 degrees, four feature quantities are obtained through calculation, and the smaller the correlation coefficient of the features is, the smaller the redundancy quantity is. The entropy feature quantity with less information redundancy is selected, is a measure of randomness and disorder of image contents, and reflects the disorder degree of textures, and the image corresponding to the feature quantity is shown in fig. 3(b), wherein the black part is a ridge, and the white part is other types of ground objects. As can be seen from the feature image, the ridge has a significant feature in the image, but is not complete, and more fine and broken ridges are not identified, and a complete ridge line cannot be obtained only by texture entropy extraction. Therefore, the texture entropy characteristics and the calculation results of the experimental variation functions are overlapped, then the FCM method is adopted to carry out clustering analysis on the results, two clustering centers are arranged in the clustering process and correspond to two types of ground objects, namely buildings and non-buildings, and building areas are obtained after clustering segmentation is finished. The building area is combined with a building and a ridge mixed ground object obtained by an area growing method, ridge lines are identified, the building, the ridge and other types of ground objects are respectively mapped into three colors of gray, white and black, and an image shown in fig. 4 is synthesized. In the figure, white parts are ridges, gray parts are buildings, and black parts are other types of ground objects.
(4) Analysis of the Ridge line extraction Effect of the invention
The accuracy of the verification results of the real ridge sample and the non-ridge sample is selected, and the accuracy verification results are shown in table 1, and the overall accuracy OA of the ridge line extraction method provided by the invention in the embodiment is 90.78%.
TABLE 1 accuracy evaluation of the ridge line extraction method of the present invention
And (3) segmenting the Span image by using a classical threshold segmentation method, and comparing and analyzing the result with the result of ridge line extraction provided by the invention. The method adopts an adaptive threshold method, namely, the local gray average value of the image is calculated according to the brightness distribution of different areas of the image to determine the threshold value, so that the image is segmented, and the ridge line extraction precision of the classic threshold segmentation method is 51.33 percent and is obviously smaller than the precision of 90.78 percent. Although the threshold segmentation method has a good extraction effect on continuous wide ridges in an image, the extraction effect is poor at a plurality of discontinuous hillside ridges, and the hillside ridges with more discontinuities are not extracted, so that the extraction accuracy of the algorithm is greatly reduced. Therefore, the method can distinguish the building from the ridge, can identify the continuous large ridge and the discontinuous small ridge, and has the identification accuracy of 90.78 percent.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.
Claims (3)
1. The PolSAR image ridge line extraction method based on the region growing method and the variation function is characterized by comprising the following steps of:
(1) initial segmentation of PolSAR image by using region growing method for automatically selecting seed points
Firstly, selecting initial seed points by combining boundary information of an image, designating one seed point as a starting point of growth, merging pixels and seed points of four fields or eight fields of the initial seed points, merging all pixels with similar properties and continuing to grow outwards until pixels which do not meet conditions are added, and obtaining a mixed ground object of a ridge line and a building;
(2) calculating an omnidirectional variation function of an image to obtain variation texture features, simultaneously calculating texture entropy features of the image, superposing the variation texture features and the texture entropy features to obtain combined texture features, performing cluster analysis on the combined texture features by adopting a fuzzy C mean value clustering method, setting two clustering centers in a clustering process, respectively corresponding to two types of ground objects, namely buildings and non-buildings, and extracting building areas after completing cluster segmentation;
(3) and (3) combining the ridge line and the mixed ground feature of the building obtained in the step (1) with the building area extracted in the step (2) to obtain a ridge line identification result graph.
2. The method for extracting the ridgeline of the PolSAR image based on the region growing method and the variogram as claimed in claim 1, wherein in the step (1), the automatic selection method of the initial seed point comprises the following steps: traversing the whole image pixel by using a window with a certain size, calculating the pixel value mean value of other coordinate points except the central point in each window, calculating the difference value of the mean value and the pixel value of the central point, and selecting the coordinate points meeting the following three conditions as initial seed points at the same time, wherein the three conditions are as follows:
a. the difference calculation result of the initial seed points is close to the median of an interval formed by 0 and a threshold value of a region growing method;
b. the initial seed point falls within a ridge line region in the image;
c. the distance between the initial seed point and the central point of the map sheet is the nearest.
3. The method for extracting the ridgeline of the PolSAR image based on the region growing method and the variation function as claimed in claim 1, wherein in the step (2), the method for obtaining the omnidirectional variation function of the image comprises: taking the width of the window as W, and setting the coordinate of each pixel in the image as (i, j), wherein the coordinate of the center point of the window is (i)0,j0) Taking the average value of the calculation results of the variation functions in four directions of 0 °, 45 °, 90 ° and 135 ° as the omnidirectional variation function of the image, the calculation formula is as follows:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011206437.2A CN112241956B (en) | 2020-11-03 | 2020-11-03 | PolSAR image ridge line extraction method based on region growing method and variation function |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011206437.2A CN112241956B (en) | 2020-11-03 | 2020-11-03 | PolSAR image ridge line extraction method based on region growing method and variation function |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112241956A true CN112241956A (en) | 2021-01-19 |
| CN112241956B CN112241956B (en) | 2023-04-07 |
Family
ID=74169804
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011206437.2A Active CN112241956B (en) | 2020-11-03 | 2020-11-03 | PolSAR image ridge line extraction method based on region growing method and variation function |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112241956B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113191374A (en) * | 2021-05-19 | 2021-07-30 | 甘肃省地震局（中国地震局兰州地震研究所） | PolSAR image ridge line extraction method based on pyramid attention network |
| CN115641442A (en) * | 2022-12-02 | 2023-01-24 | 耕宇牧星(北京)空间科技有限公司 | Mountain image ridge line segmentation method based on fusion network |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040130546A1 (en) * | 2003-01-06 | 2004-07-08 | Porikli Fatih M. | Region growing with adaptive thresholds and distance function parameters |
| CN105205816A (en) * | 2015-09-15 | 2015-12-30 | 中国测绘科学研究院 | Method for extracting high-resolution SAR image building zone through multi-feature weighted fusion |
| CN108830203A (en) * | 2018-06-01 | 2018-11-16 | 中国科学院遥感与数字地球研究所 | A kind of building area extracting method and system based on SAR image |
| WO2019000653A1 (en) * | 2017-06-30 | 2019-01-03 | 清华大学深圳研究生院 | Image target identification method and apparatus |
| CN110147795A (en) * | 2019-05-24 | 2019-08-20 | 电子科技大学 | A kind of adaptive non local fuzzy C-means clustering SAR image partitioning algorithm |
| CN110349160A (en) * | 2019-06-25 | 2019-10-18 | 电子科技大学 | One kind is based on super-pixel and fuzzy C-means clustering SAR image segmentation method |
-
2020
- 2020-11-03 CN CN202011206437.2A patent/CN112241956B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040130546A1 (en) * | 2003-01-06 | 2004-07-08 | Porikli Fatih M. | Region growing with adaptive thresholds and distance function parameters |
| CN105205816A (en) * | 2015-09-15 | 2015-12-30 | 中国测绘科学研究院 | Method for extracting high-resolution SAR image building zone through multi-feature weighted fusion |
| WO2019000653A1 (en) * | 2017-06-30 | 2019-01-03 | 清华大学深圳研究生院 | Image target identification method and apparatus |
| CN108830203A (en) * | 2018-06-01 | 2018-11-16 | 中国科学院遥感与数字地球研究所 | A kind of building area extracting method and system based on SAR image |
| CN110147795A (en) * | 2019-05-24 | 2019-08-20 | 电子科技大学 | A kind of adaptive non local fuzzy C-means clustering SAR image partitioning algorithm |
| CN110349160A (en) * | 2019-06-25 | 2019-10-18 | 电子科技大学 | One kind is based on super-pixel and fuzzy C-means clustering SAR image segmentation method |
Non-Patent Citations (3)
| Title |
|---|
| 林晨曦等: "基于变差函数的中高分辨率SAR影像农村建筑区提取", 《中国图象图形学报》 * |
| 肖修来等: "结合变差纹理特征的极化SAR建筑物震害信息提取", 《地震工程学报》 * |
| 薛寒等: "基于变差函数的SAR图像分类方法", 《现代电子技术》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113191374A (en) * | 2021-05-19 | 2021-07-30 | 甘肃省地震局（中国地震局兰州地震研究所） | PolSAR image ridge line extraction method based on pyramid attention network |
| CN113191374B (en) * | 2021-05-19 | 2023-04-18 | 甘肃省地震局（中国地震局兰州地震研究所） | PolSAR image ridge line extraction method based on pyramid attention network |
| CN115641442A (en) * | 2022-12-02 | 2023-01-24 | 耕宇牧星(北京)空间科技有限公司 | Mountain image ridge line segmentation method based on fusion network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112241956B (en) | 2023-04-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108830870B (en) | Satellite image high-precision farmland boundary extraction method based on multi-scale structure learning | |
| CN106909902B (en) | Remote sensing target detection method based on improved hierarchical significant model | |
| CN109871902B (en) | SAR small sample identification method based on super-resolution countermeasure generation cascade network | |
| CN110309781B (en) | House damage remote sensing identification method based on multi-scale spectrum texture self-adaptive fusion | |
| CN107330875B (en) | Water body surrounding environment change detection method based on forward and reverse heterogeneity of remote sensing image | |
| CN110349160B (en) | SAR image segmentation method based on super-pixel and fuzzy C-means clustering | |
| CN108830844B (en) | Facility vegetable extraction method based on multi-temporal high-resolution remote sensing image | |
| CN110598564B (en) | OpenStreetMap-based high-spatial-resolution remote sensing image transfer learning classification method | |
| CN104361589A (en) | High-resolution remote sensing image segmentation method based on inter-scale mapping | |
| CN105335966A (en) | Multi-scale remote-sensing image segmentation method based on local homogeneity index | |
| CN109829423B (en) | Infrared imaging detection method for frozen lake | |
| CN112241956B (en) | PolSAR image ridge line extraction method based on region growing method and variation function | |
| CN106611422B (en) | Stochastic gradient Bayes's SAR image segmentation method based on sketch structure | |
| Chabrier et al. | Optimization-based image segmentation by genetic algorithms | |
| CN105389799B (en) | SAR image object detection method based on sketch map and low-rank decomposition | |
| CN108805057B (en) | SAR image reservoir area detection method based on joint significance analysis | |
| CN111160127A (en) | Remote sensing image processing and detecting method based on deep convolutional neural network model | |
| CN115641327B (en) | Building engineering quality supervision and early warning system based on big data | |
| US11941878B2 (en) | Automated computer system and method of road network extraction from remote sensing images using vehicle motion detection to seed spectral classification | |
| CN108734122B (en) | Hyperspectral urban water body detection method based on self-adaptive sample selection | |
| CN110570462A (en) | flood inundation range automatic extraction method based on polarized radar remote sensing image | |
| CN115690086A (en) | Object-based high-resolution remote sensing image change detection method and system | |
| Zheng et al. | Building recognition of UAV remote sensing images by deep learning | |
| CN114266947A (en) | Classification method and device based on fusion of laser point cloud and visible light image | |
| CN107292268A (en) | The SAR image semantic segmentation method of quick ridge ripple deconvolution Structure learning model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |