WO2022041598A1 - 一种遥感影像分割方法、系统、终端以及存储介质 - Google Patents

一种遥感影像分割方法、系统、终端以及存储介质 Download PDF

Info

Publication number
WO2022041598A1
WO2022041598A1 PCT/CN2020/139238 CN2020139238W WO2022041598A1 WO 2022041598 A1 WO2022041598 A1 WO 2022041598A1 CN 2020139238 W CN2020139238 W CN 2020139238W WO 2022041598 A1 WO2022041598 A1 WO 2022041598A1
Authority
WO
WIPO (PCT)
Prior art keywords
fuzzy
remote sensing
sub
pixel
sensing image
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.)
Ceased
Application number
PCT/CN2020/139238
Other languages
English (en)
French (fr)
Inventor
李晓丽
陈劲松
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Institute of Advanced Technology of CAS
Original Assignee
Shenzhen Institute of Advanced Technology of CAS
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shenzhen Institute of Advanced Technology of CAS filed Critical Shenzhen Institute of Advanced Technology of CAS
Publication of WO2022041598A1 publication Critical patent/WO2022041598A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/12—Edge-based segmentation
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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 OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10032—Satellite or aerial image; Remote sensing
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20021—Dividing image into blocks, subimages or windows
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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

Definitions

  • the invention belongs to the technical field of remote sensing image processing, and in particular relates to a remote sensing image segmentation method, system, terminal and storage medium.
  • Image segmentation is to divide the image into several non-overlapping homogeneous regions according to the similarity of certain predetermined features, and make these features show a high degree of consistency in the homogeneous region and obvious differences between the homogeneous regions. sex.
  • Remote sensing image segmentation is one of the key steps in remote sensing image interpretation, and the accuracy of segmentation will directly affect the quality of subsequent image interpretation.
  • the fuzzy clustering objective function is constructed by coupling the membership degree and the dissimilarity measure between pixels and clusters, and the membership degree is iteratively solved with the objective function minimization as the criterion, and the best segmentation result is obtained through the defuzzification operation.
  • the noise immunity is poor due to the lack of consideration of spatial information in traditional FCM algorithms.
  • the present invention provides a remote sensing image segmentation method, system, terminal and storage medium, aiming at solving the existing remote sensing image segmentation method.
  • the ability of feature features leads to the technical problem that it is difficult to achieve effective segmentation of complex boundaries.
  • the technical solutions adopted in the embodiments of the present invention include:
  • a remote sensing image segmentation method comprising the following steps:
  • the fuzzy attribution of the pixel to the divided sub-regions is calculated, and according to the fuzzy attribution, the image domain space of the remote sensing image to be divided is divided into spatial fuzzy geometrical space, and a specified number of fuzzy divisions are generated. subregion;
  • the parameters in the objective function are solved, and the segmentation result of the remote sensing image to be segmented is output.
  • the technical solution adopted in the embodiment of the present invention further includes: before calculating the fuzzy attribution degree of the divided sub-regions of the pixel pair according to the fuzzy dissimilarity between the pixel and the generated point, the method further includes:
  • F(d ij ⁇ , d ij C ) is the coupling function of spectral distance and spatial distance.
  • the technical solution adopted in the embodiment of the present invention further includes: the fuzzy attribution of the pixel pair to the divided sub-regions is:
  • ⁇ 1 is the ambiguity factor
  • the technical solution adopted in the embodiment of the present invention further includes: before calculating the fuzzy membership degree of the fuzzy sub-region to the cluster according to the fuzzy dissimilarity between the fuzzy sub-region and the cluster, the method further includes:
  • the pixel-level fuzzy dissimilarity s il is weighted with the fuzzy belonging degree r ij to calculate the fuzzy dissimilarity S jl between the fuzzy subregion j and the cluster l:
  • L (i) represents the label of the fuzzy sub-region to which pixel i belongs.
  • the technical solution adopted in the embodiment of the present invention further includes: the fuzzy membership degree of the fuzzy division sub-regions to the clustering is:
  • ⁇ 2 is the ambiguity factor
  • the technical solution adopted in the embodiment of the present invention further includes: the fuzzy clustering objective function is:
  • Fuzzy membership degree matrix G is the generated point set
  • V is the cluster center
  • a remote sensing image segmentation system comprising:
  • Spatial fuzzy geometric division module used to calculate the fuzzy attribution degree of the pixel to the divided sub-regions according to the fuzzy dissimilarity between the pixel and the generated point, and perform spatial fuzzy geometry calculation on the image domain space of the remote sensing image to be divided according to the fuzzy attribution degree. Divide, generate a specified number of fuzzy divided sub-regions;
  • Fuzzy cluster segmentation module used to calculate the fuzzy membership degree of the fuzzy sub-region to the cluster according to the fuzzy dissimilarity between the fuzzy sub-region and the cluster, and construct the fuzzy membership degree based on the fuzzy membership degree and the fuzzy membership degree
  • the fuzzy clustering objective function of the fuzzy divided sub-regions the parameters in the objective function are solved, and the segmentation result of the remote sensing image to be segmented is output.
  • a terminal includes a processor and a memory coupled to the processor, wherein,
  • the memory stores program instructions for implementing the remote sensing image segmentation method
  • the processor is configured to execute the program instructions stored in the memory to control remote sensing image segmentation.
  • Another technical solution adopted by the embodiments of the present invention is: a storage medium storing program instructions executable by a processor, where the program instructions are used to execute the remote sensing image segmentation method.
  • the beneficial effect of the embodiments of the present invention is that the remote sensing image segmentation method, system, terminal and storage medium of the embodiments of the present invention further improve the classification of clusters based on the uncertainty of the sub-regions for clustering.
  • fuzzy sub-regions and use it as the basic operation unit to construct a fuzzy expression model of image features, so as to fully consider the transitivity of pixels to the uncertainty of the attribution of fuzzy sub-regions, based on
  • the model models the objective function of fuzzy clustering, and finally solves different types of parameters in the objective function to obtain the segmentation results of remote sensing images, which improves the ability of the model to describe the features of ground objects, solves the problem of complex boundary segmentation, and further improves the segmentation accuracy of remote sensing images.
  • FIG. 1 is a flowchart of a remote sensing image segmentation method according to an embodiment of the present invention
  • FIG. 2 is a schematic structural diagram of a remote sensing image segmentation system according to an embodiment of the present invention.
  • FIG. 3 is a remote sensing image to be segmented and a template diagram thereof according to an embodiment of the present invention, wherein (a1) is an IKONOS image, (b1) is a WorldView-2 image, and (a2) and (b2) are (a1) and (b1) respectively ) of the template image;
  • Fig. 4 is the segmentation result of the two images to be segmented when m is 60 and 80 respectively by the algorithm according to the embodiment of the present invention; wherein, (a1) and (b1) are the fuzzy geometric division diagrams, and (a2) and (b2) are the segmentation images. As a result, (a3) and (b3) are contour plots, and (a4) and (b4) are overlay plots;
  • Figure 5 shows the segmentation results of the two images to be segmented by the comparison algorithms FCM algorithm, RFLICM algorithm and RFCM algorithm respectively; among them, (a1) and (d1) are the segmentation results of the FCM algorithm, (a2) and (d2) are the segmentation results of the FCM algorithm The contour plots of the results, (a3) and (d3) are the overlay plots of the segmentation results of the FCM algorithm. (b1) and (e1) are the segmentation results of the RFLICM algorithm, (b2) and (e2) are the contour maps of the segmentation results of the RFLICM algorithm, and (b3) and (e3) are the overlay images of the segmentation results of the RFLICM algorithm.
  • (c1) and (f1) are the segmentation results of the RFCM algorithm
  • (c2) and (f2) are the contour diagrams of the segmentation results of the RFCM algorithm
  • (c3) and (f3) are the overlay images of the segmentation results of the RFCM algorithm
  • FIG. 6 is a schematic structural diagram of a terminal according to an embodiment of the present invention.
  • FIG. 7 is a schematic structural diagram of a storage medium according to an embodiment of the present invention.
  • the remote sensing image segmentation method of the embodiment of the present invention constructs a fuzzy partition by further introducing the uncertainty of the attribution of pixels to the divided sub-regions on the basis of the uncertainty of the cluster classification of the divided sub-regions. area, and use it as the basic operation unit to build a fuzzy representation model of image features, model a fuzzy clustering objective function based on this model, and finally solve the different types of parameters in the objective function to obtain remote sensing image segmentation results, and improve the model’s ability to describe the features of ground objects.
  • FIG. 1 is a flowchart of a remote sensing image segmentation method according to an embodiment of the present invention.
  • the remote sensing image segmentation method of the embodiment of the present invention includes the following S:
  • initializing the parameters of the image segmentation model specifically includes:
  • S3 Perform spatial fuzzy geometric division on the image domain space, and generate several fuzzy division sub-regions
  • the spatial fuzzy geometric division includes space regular division and irregular space division.
  • the spatial fuzzy geometric division method specifically includes:
  • d[ ⁇ ] represents a distance function.
  • Euclidean distance is used as an example to model the distance function. It can be understood that in other embodiments of the present invention, Mahalanobis distance, Manhattan distance, Chebyshev distance, Minkowski distance, correlation coefficient can also be used. , probability density function and information entropy and other measures that can be used as distance measure equivalence measures.
  • F(d ij ⁇ , d ij C ) is the coupling function of spectral distance and spatial distance
  • w ⁇ is the weight of spatial distance, which is used to control the noise immunity of the fuzzy sub-regions.
  • the fuzzy dissimilarity calculation between the pixel and the generated point involves at least two or more coupling schemes with different distances.
  • a linear weighted model is used as an example for modeling, where the weighting coefficient is regarded as Constant, in other embodiments of the present invention, the weighting coefficient can also be regarded as a variable, and other forms of coupling schemes such as nonlinear weighting can be adopted.
  • the fuzzy attribution is used to characterize the uncertainty of the pixel attribution to the sub-regions. Based on the calculation result of the fuzzy attribution, the image domain space is divided into m fuzzy sub-regions.
  • calculating the fuzzy dissimilarity between the fuzzy sub-regions and the clusters specifically includes:
  • L (i) represents the label of the fuzzy sub-region to which pixel i belongs.
  • the fuzzy membership degree is used to characterize the uncertainty of the fuzzy sub-regions to the clustering category.
  • ⁇ 1 and ⁇ 2 are ambiguity factors.
  • ⁇ 1 ⁇ 2 .
  • the method for solving the objective function parameter is specifically:
  • the embodiment of the present invention adopts the derivative method to directly solve
  • the embodiment of the present invention adopts the method of constructing a Lagrangian function to solve
  • the embodiments of the present invention design a numerical approximation solution based on the objective function minimization.
  • updating model parameters specifically includes:
  • the fuzzy attribution degree r ij of the pixel to the fuzzy sub-region has constraints
  • the Lagrangian function L 1 (R, U, G, V) is constructed according to the fuzzy clustering objective function J (R, U, G, V):
  • ⁇ i is a Lagrange multiplier.
  • the fuzzy sub-region j has constraints on the fuzzy membership degree u jl of cluster l
  • the Lagrangian function L 2 (R, U, G, V) is constructed according to the fuzzy clustering objective function J (R, U, G, V):
  • ⁇ j is the Lagrange multiplier.
  • FIG. 2 is a schematic structural diagram of a remote sensing image segmentation system according to an embodiment of the present invention.
  • the remote sensing image segmentation system 40 according to the embodiment of the present invention includes:
  • the spatial fuzzy geometric division module 41 is used to calculate the fuzzy attribution degree of the pixel to the divided sub-regions according to the fuzzy dissimilarity between the pixels and the generated points, and perform spatial blurring of the image domain space of the remote sensing image to be divided according to the fuzzy attribution degree Geometric division, generating a specified number of fuzzy division sub-regions;
  • Fuzzy cluster segmentation module 42 used to calculate the fuzzy membership degree of the fuzzy partition sub-region to the cluster according to the fuzzy dissimilarity between the fuzzy partition sub-region and the cluster, based on the fuzzy membership degree and the fuzzy membership degree Constructing the fuzzy clustering objective function of the fuzzy divided sub-regions; solving the parameters in the objective function, and outputting the segmentation result of the remote sensing image to be segmented.
  • the following embodiments conduct simulation experiments through design comparison tests. Specifically: using the embodiment of the present invention and the traditional FCM (Fuzzy C-means, fuzzy C-means) algorithm, RFLICM algorithm, RFCM algorithm to segment different types of remote sensing images respectively, and use the template as the standard to segment the results of each algorithm. Quantitative evaluation.
  • FCM Fuzzy C-means, fuzzy C-means
  • Figure 3 is a schematic diagram of an image to be segmented, in which (a1) is an IKONOS image with a resolution of 1m; (b1) is a WorldView-2 image with a resolution of 0.5m; (a1) and (b1) are hereinafter referred to as two to be segmented Image; (a2) and (b2) are template images of (a1) and (b1), respectively.
  • FIG. 4 is a segmentation result of two images to be segmented when m is taken as 60 and 80 by an algorithm according to an embodiment of the present invention.
  • (a1) and (b1) are the fuzzy geometric division maps
  • (a2) and (b2) are the segmentation results
  • (a3) and (b3) are contour maps
  • (a4) and (b4) are overlay maps.
  • the embodiment of the present invention can not only overcome the complex noises in regions I and III in (a1) of FIG. 3 and region IV in (b1) in FIG.
  • the segmentation effect of features with slender and complex boundary features can be improved, and effective segmentation of the region II in (b1) of FIG. 3 can be achieved.
  • Figure 5 shows the segmentation results of the two images to be segmented by the comparison algorithms FCM algorithm, RFLICM algorithm and RFCM algorithm respectively, in which (a1) and (d1) are the segmentation results of the FCM algorithm, (a2) and (d2) are the segmentation results of the FCM algorithm The contour plots of the results, (a3) and (d3) are the overlay plots of the segmentation results of the FCM algorithm.
  • (b1) and (e1) are the segmentation results of the RFLICM algorithm
  • (b2) and (e2) are the contour maps of the segmentation results of the RFLICM algorithm
  • (b3) and (e3) are the overlay images of the segmentation results of the RFLICM algorithm.
  • (c1) and (f1) are the segmentation results of the RFCM algorithm
  • (c2) and (f2) are the contour maps of the segmentation results of the RFCM algorithm
  • (c3) and (f3) are the overlay images of the segmentation results of the RFCM algorithm. It can be seen from Figure 5 that the FCM algorithm has extremely poor noise resistance, resulting in a large amount of segmentation noise in the segmentation results of Region I in Figure 3 (a1) and Region IV in Figure 3 (b1), and the existence of homogeneous region boundaries.
  • segmentation accuracy indicators of the embodiment of the present invention are significantly higher than those of the comparison algorithm, and the total accuracy can reach up to 97.66%, and the Kappa value is 0.96, which accurately verifies the effectiveness of the embodiment of the present invention.
  • FIG. 6 is a schematic structural diagram of a terminal according to an embodiment of the present invention.
  • the terminal 50 includes a processor 51 and a memory 52 coupled to the processor 51 .
  • the memory 52 stores program instructions for implementing the above-mentioned remote sensing image segmentation method.
  • the processor 51 is configured to execute program instructions stored in the memory 52 to control remote sensing image segmentation.
  • the processor 51 may also be referred to as a CPU (Central Processing Unit, central processing unit).
  • the processor 51 may be an integrated circuit chip with signal processing capability.
  • the processor 51 may also be a general purpose processor, digital signal processor (DSP), application specific integrated circuit (ASIC), off-the-shelf programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware component .
  • DSP digital signal processor
  • ASIC application specific integrated circuit
  • FPGA off-the-shelf programmable gate array
  • a general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
  • FIG. 7 is a schematic structural diagram of a storage medium according to an embodiment of the present invention.
  • the storage medium of the embodiment of the present invention stores a program file 61 capable of implementing all the above methods, wherein the program file 61 may be stored in the above-mentioned storage medium in the form of a software product, and includes several instructions to make a computer device (which may It is a personal computer, a server, or a network device, etc.) or a processor (processor) that executes all or part of the methods of the various embodiments of the present invention.
  • a computer device which may It is a personal computer, a server, or a network device, etc.
  • processor processor
  • the aforementioned storage medium includes: U disk, mobile hard disk, Read-Only Memory (ROM, Read-Only Memory), Random Access Memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program codes , or terminal devices such as computers, servers, mobile phones, and tablets.
  • the remote sensing image segmentation method, system, terminal, and storage medium of the embodiments of the present invention further introduce the uncertainty of the belonging of pixels to the divided sub-areas on the basis of the uncertainty of the clustering category of the divided sub-areas, and construct the fuzzy divided sub-areas.
  • using the fuzzy sub-region as the basic operation unit to construct the image feature fuzzy expression model, based on the model to model the fuzzy clustering objective function, and finally solve the different types of parameters in the objective function to obtain the remote sensing image segmentation results improve the model's description of ground objects It can solve the complex boundary segmentation problem and improve the segmentation accuracy of remote sensing images.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Image Analysis (AREA)

Abstract

本发明涉及一种遥感影像分割方法、系统、终端以及存储介质。包括：根据像素与生成点间的模糊非相似性计算所述像素对划分子区域的模糊归属度，根据所述模糊归属度对待分割遥感影像的影像域空间进行空间模糊几何划分，生成指定数量的模糊划分子区域；根据所述模糊划分子区域与聚类间的模糊非相似性计算所述模糊划分子区域对聚类的模糊隶属度；基于所述模糊归属度以及模糊隶属度构建所述模糊划分子区域的模糊聚类目标函数；对所述目标函数中的参数进行求解，输出所述待分割遥感影像的分割结果。本发明提高了模型刻画地物特征的能力，解决了复杂边界的误分割问题，进而提高遥感影像分割精度。

Description

一种遥感影像分割方法、系统、终端以及存储介质 技术领域
本发明属于遥感影像处理技术领域，特别涉及一种遥感影像分割方法、系统、终端以及存储介质。
背景技术
影像分割即根据某些预定特征的相似性将影像分割成若干个互不重叠的同质区域，并使这些特征在同质区域内呈现出高度的一致性、同质区域间呈现出明显的差异性。遥感影像分割是遥感影像解译的关键步骤之一，分割的精度将直接影响后续影像解译的质量。
随着传感器分辨率的不断提高，高分辨率遥感影像提供更多丰富地物细节信息的同时，也使得影像中“同谱异物，同物异谱”的现象加剧，光谱在影像域的空间相关性更加复杂，进而导致遥感影像分割的不确定性进一步增加。模糊聚类作为解决不确定性问题的有效工具之一，被广泛研究并应用于遥感影像分割。其中，最具代表性的为模糊C均值聚类(Fuzzy C-means,FCM)算法，该算法利用隶属度构建像素关于所有同质区域的模糊隶属关系，有效地描述了分割过程中像素类属的不确定性，通过耦合隶属度与像素和聚类间非相似性测度构建模糊聚类目标函数，以目标函数最小化为准则迭代求解隶属度，并经反模糊化操作获得最佳分割结果。然而，由于传统FCM算法缺乏考虑空间信息的作用，导致抗噪性较差。
为了解决上述不足，众多学者提出了一系列融入邻域像素空间作用的模糊聚类算法，例如顾及空间约束的FCM算法(Fuzzy Clustering with Constraint,FCM_S)、鲁棒模糊局部信息C均值聚类算法(Fuzzy Local Information C-Means,FLICM)以及修正的FLICM算法(Reformulated FLICM,RFLICM)等，均通过直接在目标函数中增加具有邻域像素作用的规则化项以引入空间信息约束；此外，还包括增强FCM算法(Enhanced FCM,EnFCM)和快速广义FCM算法(Fast Generalized FCM,FGFCM)等，利用邻域像素光谱信息构建加权影像，以引入空间信息作用，然后基于加权影像实现分割。但是，上述算法均以像素为基本操作单元，克服噪声能力有限，难以解决高分辨率遥感影像中复杂细节信息对分割 结果的影响。
近几年，基于空间几何划分的遥感影像分割方法被广泛研究。以划分子区域为基本操作单元建立像素光谱测度的或然率模型，根据贝叶斯定理得到像素关于同质区域的后验概率，并利用马尔科夫链蒙特卡洛方法(Markov Chain Monte Carlo,MCMC)求解模型参数，进而实现分割，有效地解决了遥感影像中复杂噪声对分割结果的影响，极大地提高了算法的抗噪性。而后，该思想被扩展到模糊聚类领域，结合FCM算法提出了区域化模糊C均值聚类算法(Rgionalized Fuzzy C means,RFCM)，有效地描述了划分子区域对聚类的类属不确定性，进一步提高了影像分割精度。然而，上述算法均只基于光谱信息建模了划分子区域的类属不确定性，忽略了空间信息角度像素对划分子区域归属的不确定性，限制了模型准确刻画地物特征的能力，难以实现复杂边界的有效分割。此外，当硬性划分子区域上升到模糊划分子区域时，现有模型的有效性也会受到巨大挑战。
发明内容
针对现有技术不足，本发明提供一种遥感影像分割方法、系统、终端及存储介质，旨在解决现有遥感影像分割方法存在的忽略像素对划分子区域归属的不确定性，限制模型准确刻画地物特征的能力，导致难以实现复杂边界有效分割的技术问题。
为了解决上述问题，本发明实施例采取的技术方案包括：
一种遥感影像分割方法，包括以下步骤：
根据像素与生成点间的模糊非相似性计算所述像素对划分子区域的模糊归属度，根据所述模糊归属度对待分割遥感影像的影像域空间进行空间模糊几何划分，生成指定数量的模糊划分子区域；
根据所述模糊划分子区域与聚类间的模糊非相似性计算所述模糊划分子区域对聚类的模糊隶属度；
基于所述模糊归属度以及模糊隶属度构建所述模糊划分子区域的模糊聚类目标函数；
对所述目标函数中的参数进行求解，输出所述待分割遥感影像的分割结果。
本发明实施例采取的技术方案还包括：所述根据像素与生成点间的模糊非相似性计算所述像素对划分子区域的模糊归属度之前还包括：
在所述影像域空间Ω中选取m个格点构成初始生成点集G＝{g j(a j,b j):(a j,b j)∈Ω,j＝1,...,m}，其中j为生成点索引，m为生成点数，g j为生成点j的光谱测度矢量，g j＝(g js:s＝1,...,h)，s为波段索引，h为波段数，(a j,b j)为生成点j在影像空间域的格点位置坐标；
计算像素i与生成点j间的光谱距离d ij C＝d[I i,g j]，其中，d[·]表示距离函数，I i为像素i在特征域的光谱测度矢量，I i＝(I is:s＝1,...,h)，；
计算像素i与生成点j间的空间距离d ij Ω＝d[(x i,y i),(a j,b j)]，其中，(x i,y i)为像素i在影像空间域的格点位置坐标；
耦合所述光谱距离和所述空间距离得到像素i与生成点j间的模糊非相似性d ij：
Figure PCTCN2020139238-appb-000001
上式中，F(d ij Ω,d ij C)为光谱距离和空间距离耦合函数。
本发明实施例采取的技术方案还包括：所述像素对划分子区域的模糊归属度为：
Figure PCTCN2020139238-appb-000002
上式中，λ 1为模糊因子。
本发明实施例采取的技术方案还包括：所述根据所述模糊划分子区域与聚类间的模糊非相似性计算所述模糊划分子区域对聚类的模糊隶属度之前还包括：
计算像素i与聚类l间的像素级模糊非相似性s il：
s il＝d[I i,v l]
上式中，v l为第l个聚类的聚类中心，v l＝(v ls:s＝1,...,h)；
以所述模糊归属度r ij加权所述像素级模糊非相似性性s il计算模糊划分子区域j与聚类l间的模糊非相似性S jl：
Figure PCTCN2020139238-appb-000003
上式中，L (i)表示像素i归属模糊划分子区域的标号。
本发明实施例采取的技术方案还包括：所述模糊划分子区域对聚类的模糊隶属度为：
Figure PCTCN2020139238-appb-000004
上式中，λ 2为模糊因子。
本发明实施例采取的技术方案还包括：所述模糊聚类目标函数为：
Figure PCTCN2020139238-appb-000005
上式中，R＝[r ij] n×m为所述像素关于所述模糊划分子区域的模糊归属度矩阵；U＝[u jl] m×k为所述模糊划分子区域关于聚类的模糊隶属度矩阵；G为生成点集，V为聚类中心，V＝{v l:l＝1,...,k}，其中l为聚类索引，k为聚类数。
本发明实施例采取的另一技术方案为：一种遥感影像分割系统，包括：
空间模糊几何划分模块：用于根据像素与生成点间的模糊非相似性计算所述像素对划分子区域的模糊归属度，根据所述模糊归属度对待分割遥感影像的影像域空间进行空间模糊几何划分，生成指定数量的模糊划分子区域；
模糊聚类分割模块：用于根据所述模糊划分子区域与聚类间的模糊非相似性计算所述模糊划分子区域对聚类的模糊隶属度，基于所述模糊归属度以及模糊隶属度构建所述模糊划分子区域的模糊聚类目标函数；对所述目标函数中的参数进行求解，输出所述待分割遥感影像的分割结果。
本发明实施例采取的又一技术方案为：一种终端，所述终端包括处理器、与所述处理器耦接的存储器，其中，
所述存储器存储有用于实现所述遥感影像分割方法的程序指令；
所述处理器用于执行所述存储器存储的所述程序指令以控制遥感影像分割。
本发明实施例采取的又一技术方案为：一种存储介质，存储有处理器可运行的程序指令，所述程序指令用于执行所述遥感影像分割方法。
相对于现有技术，本发明实施例产生的有益效果在于：本发明实施例的遥感影像分割方法、系统、终端以及存储介质通过在划分子区域对聚类类属不确定性的基础上，进一步引入像素对划分子区域归属的不确定性，构建模糊划分子区域，并以其作为基本操作单元构建影像特征模糊表达模型，以充分考虑像素对模糊划分子区域归属不确定性的传递性，基于该模型建模模糊聚类目标函数，最后对目标函数中不同类型的参数进行求解得到遥感 影像分割结果，提高模型刻画地物特征的能力，解决复杂边界分割问题，进而提高遥感影像分割精度。
附图说明
图1是本发明实施例的遥感影像分割方法的流程图；
图2为本发明实施例的遥感影像分割系统结构示意图；
图3为本发明实施例的待分割遥感影像及其模板图，其中，(a1)为IKONOS影像，(b1)为WorldView-2影像，(a2)和(b2)分别为(a1)和(b1)的模板影像；
图4为本发明实施例算法以m分别取60和80时对两个待分割影像的分割结果；其中，(a1)和(b1)为模糊几何划分图，(a2)和(b2)为分割结果，(a3)和(b3)为轮廓线图，(a4)和(b4)为叠加图；
图5为对比算法FCM算法、RFLICM算法及RFCM算法分别对两个待分割影像的分割结果；其中，(a1)和(d1)为FCM算法分割结果，(a2)和(d2)为FCM算法分割结果的轮廓线图，(a3)和(d3)为FCM算法分割结果的叠加图。(b1)和(e1)为RFLICM算法分割结果，(b2)和(e2)为RFLICM算法分割结果的轮廓线图，(b3)和(e3)为RFLICM算法分割结果的叠加图。(c1)和(f1)为RFCM算法分割结果，(c2)和(f2)为RFCM算法分割结果的轮廓线图，(c3)和(f3)为RFCM算法分割结果的叠加图；
图6为本发明实施例的终端结构示意图；
图7为本发明实施例的存储介质的结构示意图。
具体实施方式
为了使本发明的目的、技术方案及优点更加清楚明白，以下结合附图及实施例，对本发明进行进一步详细说明。应当理解，此处所描述的具体实施例仅用以解释本发明，并不用于限定本发明。
针对现有技术的不足，本发明实施例的遥感影像分割方法通过在划分子区域对聚类类属不确定性的基础上，进一步引入像素对划分子区域归属的不确定性，构建模糊划分子区域，并以其作为基本操作单元构建影像特征模糊表达模型，基于该模型建模模糊聚类目标函数，最后对目标函数中不同类型的参数进行求解得到遥感影像分割结果，提高模型刻画地物特征的能力，解决复杂边界分割问题，进而提高遥感影像分割精度。
具体的，请参阅图1，是本发明实施例的遥感影像分割方法的流程图。本发明实施例的遥感影像分割方法包括以下S：
S1：通过影像分割模型读取待分割遥感影像；
本步骤中，给定一幅高分辨率遥感影像I＝{I i(x i,y i):(x i,y i)∈Ω,i＝1,...,n}，其中i为像素索引，(x i,y i)为像素i在影像空间域的格点位置坐标，I i为像素i在特征域的光谱测度矢量，I i＝(I is:s＝1,...,h)，s为波段索引，h为波段数，n为总像素数，Ω为影像域空间，Ω＝{(x i,y i):i＝1,...,n}。
S2：初始化影像分割模型参数；
本步骤中，初始化影像分割模型参数具体包括：
S2.1：初始化生成点集G；在影像域空间Ω中随机选取m个格点构成初始生成点集G＝{g j(a j,b j):(a j,b j)∈Ω,j＝1,...,m}，其中j为生成点索引，m为生成点数，由于生成点与模糊划分子区域一一对应，因此j和m又可分别称为模糊划分子区域索引和模糊划分子区域数。(a j,b j)为生成点j在影像空间域的格点位置坐标，g j为生成点j的光谱测度矢量，g j＝(g js:s＝1,...,h)。
S2.2：初始化聚类中心V；假设将遥感影像分割为k个同质区域(又称聚类)，初始化聚类中心V＝{v l:l＝1,...,k}，其中l为聚类索引，k为聚类数，v l为第l个聚类的聚类中心v l＝(v ls:s＝1,...,h)。
S3：对影像域空间进行空间模糊几何划分，生成若干个模糊划分子区域；
本步骤中，空间模糊几何划分包括空间规则划分与非规则划分，以非规则划分为例，空间模糊几何划分方式具体包括：
S3.1：计算像素i与生成点j间的光谱距离d ij C：
Figure PCTCN2020139238-appb-000006
式(1)中，d[·]表示距离函数。本发明实施例以欧氏距离为例进行距离函数的建模，可以理解，在本发明其他实施例中，也可采用马氏距离、曼哈顿距离、切比雪夫距离、闵可夫斯基距离、相关系数、概率密度函数及信息熵等其它可作为距离测度等效性的度量。
S3.2：计算像素i与生成点j间的空间距离d ij Ω：
Figure PCTCN2020139238-appb-000007
S3.3：通过耦合光谱距离和空间距离计算结果得到像素i与生成点j间的模糊非相似性d ij：
Figure PCTCN2020139238-appb-000008
式(3)中，F(d ij Ω,d ij C)为光谱距离和空间距离耦合函数，w Ω为空间距离权重，用于控制模糊划分子区域的抗噪性。为了充分发挥空间距离和光谱距离的综合作用，d ij Ω和d ij C需为归一化处理后的值。
本发明实施例中，像素与生成点间的模糊非相似性计算涉及至少两种或两种以上不同距离的耦合方案，本发明实施例以线性加权模型为例进行建模，其中加权系数视为常量，在本发明其他实施例中，也可将加权系数视为变量，以及采用非线性加权等其它形式的耦合方案。
S3.4：根据模糊非相似性d ij计算像素i对模糊划分子区域j的模糊归属度r ij：
Figure PCTCN2020139238-appb-000009
模糊归属度用于表征像素对划分子区域归属的不确定性，基于该模糊归属度计算结果，将影像域空间划分为m个模糊划分子区域
Figure PCTCN2020139238-appb-000010
S4：计算模糊划分子区域与聚类间的模糊非相似性(即影像特征模糊表达模型)；
本步骤中，计算模糊划分子区域与聚类间的模糊非相似性具体包括：
S4.1：计算像素i与聚类l间的像素级模糊非相似性s il：
Figure PCTCN2020139238-appb-000011
S4.2：以模糊归属度r ij加权像素级模糊非相似性性s il计算模糊划分子区域j与聚类l间的模糊非相似性S jl，用以刻画同一地物特征一致性：
Figure PCTCN2020139238-appb-000012
式(6)中，L (i)表示像素i归属模糊划分子区域的标号。
S5：计算各个模糊划分子区域对聚类的模糊隶属度u jl；
Figure PCTCN2020139238-appb-000013
该模糊隶属度用于表征模糊划分子区域对聚类类属的不确定性。
S6：基于模糊归属度及模糊隶属度构建模糊划分子区域的模糊聚类目标函数J(R,U,G,V)：
Figure PCTCN2020139238-appb-000014
式(8)中，R＝[r ij] n×m为像素关于模糊划分子区域的模糊归属度矩阵；U＝[u jl] m×k为模糊划分子区域关于聚类的模糊隶属度矩阵，λ 1和λ 2为模糊因子。为了便于计算，本发明实施例令λ 1＝λ 2。
本发明实施例中，目标函数参数求解方法具体为：
对于聚类中心等一般参数，本发明实施例采用导数法直接求解；
对于模糊归属度或模糊隶属度等具有约束条件的参数，本发明实施例采用构建拉格朗日函数的方式进行求解；
对于无法获取解析解的参数(例如当像素对划分子区域的非相似性采样Gamma分布等其它复杂分布时)，本发明实施例设计以目标函数最小化为准则的数值逼近求解方案。
S7：判断目标函数值是否最小化，如果否，执行S8；如果是，执行S9；
S8：更新分割模型参数，并返回S3继续执行迭代操作；
本步骤中，更新模型参数具体包括：
S8.1：更新生成点集G；令J(R,U,G,V)一阶导数为0求解，即：
Figure PCTCN2020139238-appb-000015
Figure PCTCN2020139238-appb-000016
Figure PCTCN2020139238-appb-000017
解得：
Figure PCTCN2020139238-appb-000018
Figure PCTCN2020139238-appb-000019
Figure PCTCN2020139238-appb-000020
S8.2：更新聚类中心V；令J(R,U,G,V)对v l求偏导，并令导数为0：
Figure PCTCN2020139238-appb-000021
解得：
Figure PCTCN2020139238-appb-000022
S9：停止迭代，输出遥感影像分割结果，即|J(R,U,G,V) (t+1)-J(R,U,G,V) (t)|<ε；其中，t为迭代指示器，ε为大于0的极小值。
进一步地，基于S6计算得到的模糊聚类目标函数J(R,U,G,V)，像素i对模糊划分子区域j的模糊归属度r ij的推导过程如下：
像素对模糊划分子区域的模糊归属度r ij具有约束条件
Figure PCTCN2020139238-appb-000023
由此根据模糊聚类目标函数J(R,U,G,V)构建拉格朗日函数L 1(R,U,G,V)：
Figure PCTCN2020139238-appb-000024
式(17)中，δ i为拉格朗日乘子。
基于L 1(R,U,G,V)对r ij求偏导，并令导数为0，即，
Figure PCTCN2020139238-appb-000025
根据约束条件消除拉格朗日乘子，解得r ij的解析式，
Figure PCTCN2020139238-appb-000026
同理，模糊划分子区域对聚类的模糊隶属度推导过程如下：
模糊划分子区域j对聚类l的模糊隶属度u jl具有约束条件
Figure PCTCN2020139238-appb-000027
由此根据模糊聚类目标函数J(R,U,G,V)构建拉格朗日函数L 2(R,U,G,V)：
Figure PCTCN2020139238-appb-000028
式(20)中，ξ j为拉格朗日乘子。
基于L 2(R,U,G,V)对u jl求偏导，并令导数为0，即：
Figure PCTCN2020139238-appb-000029
根据约束条件消除拉格朗日乘子，解得u jl的解析式,
Figure PCTCN2020139238-appb-000030
请参阅图2，是本发明实施例的遥感影像分割系统的结构示意图。本发明实施例的遥感影像分割系统40包括：
空间模糊几何划分模块41：用于根据像素与生成点间的模糊非相似性计算所述像素对划分子区域的模糊归属度，根据所述模糊归属度对待分割遥感影像的影像域空间进行空间模糊几何划分，生成指定数量的模糊划分子区域；
模糊聚类分割模块42：用于根据所述模糊划分子区域与聚类间的模糊非相似性计算所述模糊划分子区域对聚类的模糊隶属度，基于所述模糊归属度以及模糊隶属度构建所述模糊划分子区域的模糊聚类目标函数；对所述目标函数中的参数进行求解，输出所述待分割遥感影像的分割结果。
为了验证本发明实施例的可行性和有效性，以下实施例通过设计对比测试进行仿真实验。具体为：应用本发明实施例和传统FCM(Fuzzy C-means，模糊C均值)算法、RFLICM算法、RFCM算法分别对不同类型的遥感影像进行分割，并以模板为标准对各算法的分割结果进行定量评价。图3为待分割影像示意图，其中(a1)为IKONOS影像，分辨率为1m；(b1)为WorldView-2影像，分辨率为0.5m；(a1)和(b1)以下简称为两个待分割影像；(a2)和(b2)分别为(a1)和(b1)的模板影像。
图4为本发明实施例算法以m分别取60和80时对两个待分割影像的分割结果。其中，(a1)和(b1)为模糊几何划分图，(a2)和(b2)为分割结果，(a3)和(b3)为轮廓线图，(a4)和(b4)为叠加图。由图4可以看出，本发明实施例不仅能够克服图3的(a1)中区域I、III及图3的(b1)中区域IV的复杂噪声，实现同质区域边界的有效拟合，还可改善具有细长型复杂边界特征地物的分割效果，实现图3的(b1)中区域II的有效分割。
图5为对比算法FCM算法、RFLICM算法及RFCM算法分别对两个待分割影像的分割结果，其中，(a1)和(d1)为FCM算法分割结果，(a2)和(d2)为FCM算法分割结果的轮廓线图，(a3)和(d3)为FCM算法分割结果的叠加图。(b1)和(e1)为RFLICM算法分割结果，(b2)和(e2)为RFLICM算法分割结果的轮廓线图，(b3)和(e3)为RFLICM算法分割结果的叠加图。(c1)和(f1)为RFCM算法分割结果，(c2)和(f2)为RFCM算法分割结果的轮廓线图，(c3)和(f3)为RFCM算法分割结果的叠加图。由图5可以看出，FCM算法抗噪性极差，导致图3的(a1)中区域I和图3的(b1)中区域IV的分割结果中包含大量分割噪声，并且同质区域边界存在大量误分像素；RFLICM算法因引入邻域像素作用，部分地解决了分割噪声问题，但是边界分割效果依然较差；RFCM算法以硬性划分子区域为基本操作单元，有效地解决了同质区域内分割噪声问题，但是受硬性划分子区域影响，依然难以有效拟合复杂边界，导致图3的(b1)中区域II和IV分割混乱。
以两个待分割影像为分割模板，对本发明实施例算法和对比算法的分割结果生成混淆矩阵，并计算用户精度、产品精度、总精度和kappa值，具体的定量评价表如下表1所示：
表1本发明算法与各对比算法的定量评价
Figure PCTCN2020139238-appb-000031
Figure PCTCN2020139238-appb-000032
可以看出，本发明实施例的各项分割精度指标明显高于对比算法，并且总精度最高可达97.66％，Kappa值为0.96，精确地验证了本发明实施例的有效性。
请参阅图6，为本发明实施例的终端结构示意图。该终端50包括处理器51、与处理器51耦接的存储器52。
存储器52存储有用于实现上述遥感影像分割方法的程序指令。
处理器51用于执行存储器52存储的程序指令以控制遥感影像分割。
其中，处理器51还可以称为CPU(Central Processing Unit，中央处理单元)。处理器51可能是一种集成电路芯片，具有信号的处理能力。处理器51还可以是通用处理器、数字信号处理器(DSP)、专用集成电路(ASIC)、现成可编程门阵列(FPGA)或者其他可编程 逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。
请参阅图7，为本发明实施例的存储介质的结构示意图。本发明实施例的存储介质存储有能够实现上述所有方法的程序文件61，其中，该程序文件61可以以软件产品的形式存储在上述存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)或处理器(processor)执行本发明各个实施方式方法的全部或部分S。而前述的存储介质包括：U盘、移动硬盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质，或者是计算机、服务器、手机、平板等终端设备。
本发明实施例的遥感影像分割方法、系统、终端以及存储介质在划分子区域对聚类类属不确定性的基础上，进一步引入像素对划分子区域归属的不确定性，构建模糊划分子区域，以模糊划分子区域作为基本操作单元构建影像特征模糊表达模型，基于该模型建模模糊聚类目标函数，最后对目标函数中不同类型的参数进行求解得到遥感影像分割结果，提高模型刻画地物特征的能力，解决复杂边界分割问题，进而提高遥感影像分割精度。
对所公开的实施例的上述说明，使本领域专业技术人员能够实现或使用本发明。对这些实施例的多种修改对本领域的专业技术人员来说将是显而易见的，本发明中所定义的一般原理可以在不脱离本发明的精神或范围的情况下，在其它实施例中实现。因此，本发明将不会被限制于本发明所示的这些实施例，而是要符合与本发明所公开的原理和新颖特点相一致的最宽的范围。

Claims (9)

  1. 一种遥感影像分割方法，其特征在于，包括以下步骤：
    根据像素与生成点间的模糊非相似性计算所述像素对划分子区域的模糊归属度，根据所述模糊归属度对待分割遥感影像的影像域空间进行空间模糊几何划分，生成指定数量的模糊划分子区域；
    根据所述模糊划分子区域与聚类间的模糊非相似性计算所述模糊划分子区域对聚类的模糊隶属度；
    基于所述模糊归属度以及模糊隶属度构建所述模糊划分子区域的模糊聚类目标函数；
    对所述目标函数中的参数进行求解，输出所述待分割遥感影像的分割结果。
  2. 根据权利要求1所述的遥感影像分割方法，其特征在于，所述根据像素与生成点间的模糊非相似性计算所述像素对划分子区域的模糊归属度之前还包括：
    在所述影像域空间Ω中选取m个格点构成初始生成点集G＝{g j(a j,b j):(a j,b j)∈Ω,j＝1,...,m}，其中j为生成点索引，m为生成点数，g j为生成点j的光谱测度矢量，g j＝(g js:s＝1,...,h)，s为波段索引，h为波段数，(a j,b j)为生成点j在影像空间域的格点位置坐标；
    计算像素i与生成点j间的光谱距离d ij C＝d[I i,g j]，其中，d[·]表示距离函数，I i为像素i在特征域的光谱测度矢量，I i＝(I is:s＝1,...,h)；
    计算像素i与生成点j间的空间距离d ij Ω＝d[(x i,y i),(a j,b j)]，其中，(x i,y i)为像素i在影像空间域的格点位置坐标；
    耦合所述光谱距离和所述空间距离得到像素i与生成点j间的模糊非相似性d ij：
    Figure PCTCN2020139238-appb-100001
    上式中，F(d ij Ω,d ij C)为光谱距离和空间距离耦合函数。
  3. 根据权利要求2所述的遥感影像分割方法，其特征在于，所述像素对划分子区域的模糊归属度为：
    Figure PCTCN2020139238-appb-100002
    上式中，λ 1为模糊因子。
  4. 根据权利要求2所述的遥感影像分割方法，其特征在于，所述根据所述模糊划分子区域与聚类间的模糊非相似性计算所述模糊划分子区域对聚类的模糊隶属度之前还包括：
    计算像素i与聚类l间的像素级模糊非相似性s il：
    s il＝d[I i,v l]
    上式中，v l为第l个聚类的聚类中心，v l＝(v ls:s＝1,...,h)；
    以所述模糊归属度r ij加权所述像素级模糊非相似性性s il计算模糊划分子区域j与聚类l间的模糊非相似性S jl：
    Figure PCTCN2020139238-appb-100003
    上式中，L (i)表示像素i归属模糊划分子区域的标号。
  5. 根据权利要求4所述的遥感影像分割方法，其特征在于，所述模糊划分子区域对聚类的模糊隶属度为：
    Figure PCTCN2020139238-appb-100004
    上式中，λ 2为模糊因子。
  6. 根据权利要求1所述的遥感影像分割方法，其特征在于，所述模糊聚类目标函数为：
    Figure PCTCN2020139238-appb-100005
    上式中，R＝[r ij] n×m为所述像素关于所述模糊划分子区域的模糊归属度矩阵；U＝[u jl] m×k为所述模糊划分子区域关于聚类的模糊隶属度矩阵；G为生成点集，V为聚类中心，V＝{v l:l＝1,...,k}，其中l为聚类索引，k为聚类数。
  7. 一种遥感影像分割系统，其特征在于，包括：
    空间模糊几何划分模块：用于根据像素与生成点间的模糊非相似性计算所述像素对划分子区域的模糊归属度，根据所述模糊归属度对待分割遥感影像的影像域空间进行空间模糊几何划分，生成指定数量的模糊划分子区域；
    模糊聚类分割模块：用于根据所述模糊划分子区域与聚类间的模糊非相似性计算所述模糊划分子区域对聚类的模糊隶属度，基于所述模糊归属度以及模糊隶属度构建所述模糊划分子区域的模糊聚类目标函数；对所述目标函数中的参数进行求解，输出所述待分割遥感影像的分割结果。
  8. 一种终端，其特征在于，所述终端包括处理器、与所述处理器耦接的存储器，其中，
    所述存储器存储有用于实现权利要求1-6任一项所述的遥感影像分割方法的程序指令；
    所述处理器用于执行所述存储器存储的所述程序指令以控制遥感影像分割。
  9. 一种存储介质，其特征在于，存储有处理器可运行的程序指令，所述程序指令用于执行权利要求1至6任一项所述遥感影像分割方法。
PCT/CN2020/139238 2020-08-24 2020-12-25 一种遥感影像分割方法、系统、终端以及存储介质 Ceased WO2022041598A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010856484.5 2020-08-24
CN202010856484.5A CN112116610B (zh) 2020-08-24 2020-08-24 一种遥感影像分割方法、系统、终端以及存储介质

Publications (1)

Publication Number Publication Date
WO2022041598A1 true WO2022041598A1 (zh) 2022-03-03

Family

ID=73804367

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/139238 Ceased WO2022041598A1 (zh) 2020-08-24 2020-12-25 一种遥感影像分割方法、系统、终端以及存储介质

Country Status (2)

Country Link
CN (1) CN112116610B (zh)
WO (1) WO2022041598A1 (zh)

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115016289A (zh) * 2022-07-11 2022-09-06 国网江苏省电力有限公司 一种基于区间二型模糊的t-s模糊模型非线性系统辨识方法
CN115952432A (zh) * 2022-12-21 2023-04-11 四川大学华西医院 一种基于糖尿病数据的无监督聚类方法
CN116051976A (zh) * 2022-11-23 2023-05-02 河南理工大学 一种融合高程信息的遥感影像的处理方法
CN116070887A (zh) * 2023-04-06 2023-05-05 平原县自然资源服务中心 一种土地测绘数据智能分析管理系统
CN116129120A (zh) * 2023-02-13 2023-05-16 北京理工大学 一种联合分割视网膜眼底图像中视杯视盘的方法
CN116309575A (zh) * 2023-05-19 2023-06-23 济宁众达利电气设备有限公司 基于图像处理的电插头生产质量检测方法
CN116381607A (zh) * 2023-04-11 2023-07-04 哈尔滨工程大学 多目标击水声特征关联方法
WO2022258077A3 (zh) * 2022-07-04 2023-09-21 清华大学 基于ii型模糊粗糙模型的遥感影像特征离散化方法、装置、存储介质、计算机设备
CN117173102A (zh) * 2023-08-04 2023-12-05 航天恒星科技有限公司 基于多源数据融合的遥感数据稳定性分析方法、电子设备及存储介质
CN117746240A (zh) * 2023-12-20 2024-03-22 同济大学 一种基于混合核空间紧密度距离的模糊c均值解混方法
CN118154632A (zh) * 2024-03-22 2024-06-07 山东大学 基于智能驾驶的目标特征智能分割方法及系统

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112116610B (zh) * 2020-08-24 2024-02-27 中国科学院深圳先进技术研究院 一种遥感影像分割方法、系统、终端以及存储介质
CN114283332B (zh) * 2021-12-15 2024-07-19 深圳先进技术研究院 一种模糊聚类遥感影像分割方法、系统、终端及存储介质
CN114219822B (zh) * 2021-12-15 2024-06-21 深圳先进技术研究院 面向对象模糊聚类图像分割方法、系统、终端及存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102005034A (zh) * 2010-12-01 2011-04-06 南京大学 基于区域聚类的遥感影像分割方法
CN106373129A (zh) * 2016-09-20 2017-02-01 辽宁工程技术大学 一种基于双隶属度的fcm遥感图像分割方法
US10115197B1 (en) * 2017-06-06 2018-10-30 Imam Abdulrahman Bin Faisal University Apparatus and method for lesions segmentation
CN112116610A (zh) * 2020-08-24 2020-12-22 中国科学院深圳先进技术研究院 一种遥感影像分割方法、系统、终端以及存储介质

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8233712B2 (en) * 2006-07-28 2012-07-31 University Of New Brunswick Methods of segmenting a digital image
CN105590325B (zh) * 2016-02-22 2018-01-30 辽宁工程技术大学 基于模糊化高斯隶属函数的高分辨率遥感影像分割方法
CN106127784B (zh) * 2016-07-01 2018-08-07 辽宁工程技术大学 一种高分辨率遥感影像分割方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102005034A (zh) * 2010-12-01 2011-04-06 南京大学 基于区域聚类的遥感影像分割方法
CN106373129A (zh) * 2016-09-20 2017-02-01 辽宁工程技术大学 一种基于双隶属度的fcm遥感图像分割方法
US10115197B1 (en) * 2017-06-06 2018-10-30 Imam Abdulrahman Bin Faisal University Apparatus and method for lesions segmentation
CN112116610A (zh) * 2020-08-24 2020-12-22 中国科学院深圳先进技术研究院 一种遥感影像分割方法、系统、终端以及存储介质

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
ZHAO QUANHUA, LI HONG-YING;LI YU: "Regionalized fuzzy C-Means Algorithm for Segmentation of Color Remote Sensing Image", CONTROL AND DECISION, DONGBEI DAXUE, SHENYANG, CN, vol. 30, no. 9, 30 September 2015 (2015-09-30), CN , pages 1706 - 1710, XP055903452, ISSN: 1001-0920, DOI: 10.13195/j.kzyjc.2014.0848 *

Cited By (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2022258077A3 (zh) * 2022-07-04 2023-09-21 清华大学 基于ii型模糊粗糙模型的遥感影像特征离散化方法、装置、存储介质、计算机设备
CN115016289A (zh) * 2022-07-11 2022-09-06 国网江苏省电力有限公司 一种基于区间二型模糊的t-s模糊模型非线性系统辨识方法
CN116051976A (zh) * 2022-11-23 2023-05-02 河南理工大学 一种融合高程信息的遥感影像的处理方法
CN116051976B (zh) * 2022-11-23 2023-09-19 河南理工大学 一种融合高程信息的遥感影像的处理方法
CN115952432A (zh) * 2022-12-21 2023-04-11 四川大学华西医院 一种基于糖尿病数据的无监督聚类方法
CN115952432B (zh) * 2022-12-21 2024-03-12 四川大学华西医院 一种基于糖尿病数据的无监督聚类方法
CN116129120A (zh) * 2023-02-13 2023-05-16 北京理工大学 一种联合分割视网膜眼底图像中视杯视盘的方法
CN116070887A (zh) * 2023-04-06 2023-05-05 平原县自然资源服务中心 一种土地测绘数据智能分析管理系统
CN116381607A (zh) * 2023-04-11 2023-07-04 哈尔滨工程大学 多目标击水声特征关联方法
CN116381607B (zh) * 2023-04-11 2023-10-27 哈尔滨工程大学 多目标击水声特征关联方法
CN116309575A (zh) * 2023-05-19 2023-06-23 济宁众达利电气设备有限公司 基于图像处理的电插头生产质量检测方法
CN117173102A (zh) * 2023-08-04 2023-12-05 航天恒星科技有限公司 基于多源数据融合的遥感数据稳定性分析方法、电子设备及存储介质
CN117746240A (zh) * 2023-12-20 2024-03-22 同济大学 一种基于混合核空间紧密度距离的模糊c均值解混方法
CN117746240B (zh) * 2023-12-20 2025-08-08 同济大学 一种基于混合核空间紧密度距离的模糊c均值解混方法
CN118154632A (zh) * 2024-03-22 2024-06-07 山东大学 基于智能驾驶的目标特征智能分割方法及系统

Also Published As

Publication number Publication date
CN112116610A (zh) 2020-12-22
CN112116610B (zh) 2024-02-27

Similar Documents

Publication Publication Date Title
CN112116610A (zh) 一种遥感影像分割方法、系统、终端以及存储介质
Xiang et al. Superpixel generating algorithm based on pixel intensity and location similarity for SAR image classification
CN105405133B (zh) 一种遥感影像变化检测方法
WO2023116632A1 (zh) 基于时空记忆信息的视频实例分割方法和分割装置
CN104361611B (zh) 一种基于群稀疏鲁棒pca的运动目标检测方法
CN111242208A (zh) 一种点云分类方法、分割方法及相关设备
CN110363236B (zh) 空谱联合超图嵌入的高光谱图像极限学习机聚类方法
CN106295124A (zh) 利用多种图像检测技术综合分析基因子图相似概率量的方法
CN114881892B (zh) 基于ii型模糊粗糙模型的遥感影像特征离散化方法及装置
CN111223128A (zh) 目标跟踪方法、装置、设备及存储介质
CN108846845B (zh) 基于缩略图与分层模糊聚类的sar图像分割方法
CN114971802A (zh) 基于bim模型的商品推荐方法、装置、电子设备及介质
Palanivel et al. Mutifractals based multimodal 3D image registration
Wang et al. Image thresholding method based on Tsallis entropy correlation
Liu et al. Color image segmentation using nonparametric mixture models with multivariate orthogonal polynomials
CN105354845B (zh) 一种遥感影像半监督变化检测方法
CN116361670A (zh) 不完备多视图聚类方法、系统、设备以及存储介质
Wang et al. Adaptive and fast image superpixel segmentation approach
US10062187B1 (en) Systems and methods for reducing computer resources consumption to reconstruct shape of multi-object image
CN107544944B (zh) 一种基于图论的支持向量机核函数选择方法及其应用
CN112598663B (zh) 基于视觉显著性的粮食害虫检测方法和装置
Zou et al. Shape-based retrieval and analysis of 3D models using fuzzy weighted symmetrical depth images
Hao et al. Superpixel-based active contour model for unsupervised change detection from satellite images
Yuan et al. Anisotropic neighborhood searching for point cloud with sharp feature
CN109447943A (zh) 一种目标检测方法、系统及终端设备

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20951264

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20951264

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 05.07.2023

122 Ep: pct application non-entry in european phase

Ref document number: 20951264

Country of ref document: EP

Kind code of ref document: A1