Time series polarimetric sar cumulative change detection method based on similarity matrix

Abstract

The application discloses a time series polarimetric SAR cumulative change detection method based on a similarity matrix, relates to the technical field of remote sensing image change detection, and comprises the following steps: a similarity matrix of all time phase polarimetric SAR images in a time series is constructed; the similarity matrix is linearly transformed, and the maximum eigenvalue of the similarity matrix is calculated; a difference map of cumulative change of the polarimetric SAR images in the time series is calculated according to the maximum eigenvalue; and the cumulative change detection result of the polarimetric SAR images in the time series is obtained by segmenting the difference map. The time series polarimetric SAR cumulative change detection method based on the similarity matrix provided by the application does not repeatedly calculate a difference map, and compared with a traditional detection method of continuously detecting two changes in a time series, the detection precision and operation efficiency are improved.

Description

Technical Field

[0001] This application relates to the field of remote sensing image change detection technology, and in particular to a method for detecting cumulative changes in time-series polarimetric SAR based on similarity matrices. Background Technology

[0002] Synthetic Aperture Radar (SAR), as an active microwave sensor, is not limited by observation conditions such as lighting, clouds, rain, and fog, unlike optical remote sensing. It can achieve all-weather, all-day ground-to-ground observation, thus ensuring the continuity of dynamic change monitoring time, and has become an important monitoring method in the field of remote sensing.

[0003] Change detection using PolSAR remote sensing images from different time phases has been widely applied in many fields such as land cover change monitoring, forest biomass change monitoring, disaster assessment, glacier change monitoring, crop growth monitoring, and wetland monitoring.

[0004] However, current technologies are limited to detecting changes between two temporal polarimetric SAR images. In practical applications, such as flood monitoring, continuous remote sensing monitoring of the disaster area is required for a period of time after the flood occurs and until its end, obtaining PolSAR images from more than two temporal phases. During continuous monitoring, floods typically rise first and then recede. If current technologies are used to assess the flood-inundated area, the selection of which temporal phase of post-disaster imagery to detect changes between pre-disaster and post-disaster images is crucial. If the selected post-disaster imagery happens to be during the flood receding phase, this will lead to a reduction in the assessed affected area, causing numerous inconveniences for post-disaster reconstruction.

[0005] To address this problem, traditional detection methods typically first perform pairwise change detection using images from two consecutive adjacent time phases within a time series, and then fuse all the change detection results to obtain a cumulative change binary map of the entire time series. Traditional detection methods have the following drawbacks:

[0006] (1) It is easy to miss or misdetect;

[0007] (2) The calculation process is repetitive and the work efficiency is low.

[0008] Therefore, there is an urgent need for a time-series polarimetric SAR change detection method to address the aforementioned shortcomings. Summary of the Invention

[0009] (a) Purpose of application

[0010] Based on this, in order to avoid repeated detection of polarimetric SAR images of two consecutive adjacent time phases within a time series and to improve detection efficiency, this application discloses the following technical solution.

[0011] (II) Technical Solution

[0012] This application discloses a method for detecting cumulative changes in time-series polarimetric SAR based on similarity matrices, including:

[0013] Construct a similarity matrix for all temporal polarimetric SAR images within the time series;

[0014] Perform a linear transformation on the similarity matrix and calculate the maximum eigenvalue of the similarity matrix;

[0015] Calculate the difference map of cumulative changes in polarimetric SAR images over time based on the maximum eigenvalue;

[0016] The cumulative change detection results of polarimetric SAR images over time are obtained by segmenting the difference map.

[0017] In one possible implementation, the method for constructing the similarity matrix specifically includes:

[0018] Register polarimetric SAR images of the same region at N time phases, where N≧2;

[0019] The polarimetric backscattering of each pixel in a polarimetric SAR image is described by the coherence matrix T. Therefore, pixel p at a spatial location forms a coherence matrix over a time series. ;

[0020] Calculate the coherence matrix of every two temporal polarimetric SAR images within the time series. and similarity The similarity matrix R is obtained as follows:

[0021]

[0022] in, This indicates the operation of calculating the similarity of coherence matrices.

[0023] In one possible implementation, the coherence matrix of every two temporal polarization SAR images and similarity Defined as:

[0024] .

[0025] In one possible implementation, the expression for the coherence matrix T is:

[0026]

[0027] in, , , These represent complex image data for the three polarization channels: HH, VV, and HV, respectively. Represents finding the modulus of a complex number; Indicates that the image spatial window is averaged; superscript To find the conjugate of a complex number.

[0028] In one possible implementation, the method for calculating the difference map specifically includes:

[0029] The largest eigenvalue is calculated by diagonalizing the similarity matrix to perform a linear transformation.

[0030] By iterating through all pixels at spatial locations in the polarimetric SAR image, the difference is calculated based on the maximum eigenvalue. Calculate the difference plot DI:

[0031]

[0032] In this context, the subscripts P and Q represent the width and height of the polarimetric SAR image, i.e., the width and height of the difference map, respectively.

[0033] In one possible implementation, the diagonalization of the similarity matrix is ​​represented as:

[0034]

[0035] in, It is a diagonal matrix composed of real eigenvalues. , ; Column vector It is an eigenvalue The corresponding characteristic vector.

[0036] In one possible implementation, the maximum eigenvalue Represented as:

[0037] .

[0038] In one possible implementation, the degree of difference Represented as:

[0039] .

[0040] In one possible implementation, the method for obtaining the cumulative change detection result specifically includes:

[0041] Construct the Markov random field (MRF) energy function model of the difference graph;

[0042] Based on the simulated annealing global optimization algorithm, the MRF energy function is minimized, and the difference map is binary segmented to obtain the cumulative change detection results of the polarimetric SAR image in the time series.

[0043] (III) Beneficial Effects

[0044] The time-series polarimetric SAR cumulative change detection method disclosed in this application obtains the maximum eigenvalue by linearly transforming the similarity matrix within the time series to obtain a difference map that can describe the cumulative change within the time series. This method can detect the cumulative change of multiple polarimetric SAR images within the time series at once, avoiding repeated detection and reducing workload, while also avoiding missed detections and false detections, thereby improving the accuracy of the final cumulative change detection results. Attached Figure Description

[0045] The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain and illustrate this application, and should not be construed as limiting the scope of protection of this application.

[0046] Figure 1 This is a flowchart illustrating the time-series polarimetric SAR cumulative change detection method based on similarity matrices disclosed in this application.

[0047] Figure 2 This is a flowchart of the method for calculating the difference map in the time series polarimetric SAR cumulative change detection method based on similarity matrix disclosed in this application. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be described in more detail below with reference to the accompanying drawings.

[0049] The following is for reference. Figure 1 This application provides a detailed description of embodiments of the time-series polarimetric SAR cumulative change detection method based on similarity matrices disclosed in this application. For example... Figure 1 As shown, the method disclosed in this embodiment mainly includes steps S10 to S40.

[0050] S10, construct the similarity matrix of all temporal polarimetric SAR images within the time series;

[0051] The method for constructing the similarity matrix specifically includes steps S11 to S13:

[0052] S11: Register polarimetric SAR images of the same region at N time phases, where N≧2;

[0053] S12: The polarimetric backscattering of each pixel in a polarimetric SAR image is described by the coherence matrix T. Thus, the pixel p at a spatial location forms a coherence matrix within the time series in the time dimension. ;

[0054] S13: Calculate the coherence matrix of every two temporal polarimetric SAR images within the time series. and similarity The similarity matrix R is obtained as follows:

[0055]

[0056] in, This indicates the operation of calculating the similarity of coherence matrices.

[0057] Specifically, the coherence matrix of every two temporal polarimetric SAR images and similarity Defined as:

[0058]

[0059] The expression for the coherence matrix T is:

[0060]

[0061] in, , , (The subscript H indicates horizontal polarization, and V indicates vertical polarization) represents complex image data for the three polarization channels HH, VV, and HV, respectively; Represents finding the modulus of a complex number; Indicates that the image spatial window is averaged; superscript To find the conjugate of a complex number.

[0062] S20, perform a linear transformation on the similarity matrix and calculate the maximum eigenvalue of the similarity matrix;

[0063] Specifically, the maximum eigenvalue is calculated by diagonalizing the similarity matrix to linearly transform it.

[0064] The diagonalization of similar matrices is expressed as:

[0065]

[0066] in, It is a diagonal matrix composed of real eigenvalues. , ; Column vector It is an eigenvalue The corresponding characteristic vector.

[0067] Maximum eigenvalue Represented as:

[0068] .

[0069] S30, calculate the difference map of the cumulative change of the polarimetric SAR image in the time series based on the maximum eigenvalue;

[0070] Among them, such as Figure 2 As shown, the calculation method of this difference map specifically includes steps S31 to S32:

[0071] S31: Calculate the maximum eigenvalue by diagonalizing the similarity matrix to linearly transform it;

[0072] S32: Traverse each pixel at the spatial location of the polarimetric SAR image and calculate the difference based on the maximum eigenvalue. Calculate the difference plot DI:

[0073]

[0074] In this context, the subscripts P and Q represent the width and height of the polarimetric SAR image, i.e., the width and height of the difference map, respectively.

[0075] Among them, the degree of difference Represented as:

[0076] .

[0077] S40, obtain the cumulative change detection result of polarimetric SAR image in time series by segmenting the difference map.

[0078] Specifically, the method for obtaining the cumulative change detection results includes steps S41 to S42:

[0079] S41: Construct the Markov random field (MRF) energy function model of the difference graph;

[0080] S42: Minimize the MRF energy function based on the simulated annealing global optimization algorithm, perform binary segmentation on the difference map, and obtain the cumulative change detection results of the polarimetric SAR image in the time series.

[0081] In summary, the time-series polarimetric SAR cumulative change detection method proposed in this invention calculates the difference map by using the maximum eigenvalue of the similarity matrix, which can effectively characterize the cumulative changes of ground objects in the observation scene over time. This avoids the repetitive calculation method of traditional methods that often involves detecting changes in two consecutive adjacent time phases within a time series and then fusing the results, thus improving the efficiency of change detection and reducing the missed detection of small changes at a certain time point or time period in the time series, thereby improving the accuracy of cumulative change detection.

[0082] Furthermore, the difference map segmentation method based on the MRF model can effectively reduce false detections caused by speckle noise by comprehensively utilizing the neighborhood context information between pixels in the difference map, and achieves high-precision segmentation of cumulative change / non-change. The method provided can play an important role in SAR change detection and has good technical performance.

[0083] In the description of this application, it should be understood that the terms "center", "longitudinal", "lateral", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting the scope of protection of this application.

[0084] In this article, "first," "second," etc., are used only to distinguish them from each other, and do not indicate their importance or order.

[0085] The division of modules, units, or components in this document is merely a logical functional division. In actual implementation, other division methods are possible; for example, multiple modules and / or units may be combined or integrated into another system. Modules, units, and components described as separate parts may or may not be physically separate. Components shown as units may or may not be physical units; that is, they may be located in a specific location or distributed across grid cells. Therefore, some or all of the units can be selected to implement the embodiment's scheme according to actual needs.

[0086] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A time series polarimetric SAR change detection method based on similarity matrix, characterized in that, The application relates to a method for detecting the cumulative change of polarimetric SAR images in a time sequence. The method comprises the following steps: Constructing a similarity matrix of all time phases of polarimetric SAR images in a time sequence; Linearly transforming the similarity matrix to calculate the maximum eigenvalue of the similarity matrix; Calculating a difference map of the cumulative change of polarimetric SAR images in a time sequence according to the maximum eigenvalue; Obtaining the cumulative change detection result of polarimetric SAR images in a time sequence by segmenting the difference map; The method for calculating the difference map specifically comprises: The maximum eigenvalue is calculated based on the difference degree of each pixel in the spatial position of the polarimetric SAR image , and a difference map DI is calculated. where the subscripts P, Q represent the width and height of the polarimetric SAR image, i.e. the width and height of the difference map.

2. The similarity matrix based time series polarimetric SAR cumulative change detection method of claim 1, wherein, Linearly transforming the similarity matrix by diagonalizing the similarity matrix to calculate the maximum eigenvalue; The method for constructing the similarity matrix specifically comprises: The polarization backscattering of each pixel on the polarimetric SAR image is described by the coherence matrix T. The pixel p at the spatial position forms the coherence matrix in the time series in the time dimension ; Computing a coherence matrix of polarimetric SAR images for each two time phases within a time series and a similarity of a similarity matrix R is obtained: wherein, denotes a coherence matrix similarity operation.

3. The similarity matrix based time series polarimetric SAR cumulative change detection method of claim 2, wherein, The coherence matrix of each two temporal polarimetric SAR images and the similarity of is defined as: 。 4. The similarity matrix based time series polarimetric SAR cumulative change detection method of claim 2, wherein, Registering N time phases of polarimetric SAR images of the same region, wherein N>=2; wherein, , , HH, VV, HV represent the complex image data of the three polarization channels, respectively; denotes complex modulus; denotes image space windowed averaging; superscript denotes complex conjugate.

5. The similarity matrix based time series polarimetric SAR cumulative change detection method of claim 1, wherein, The expression of the coherence matrix T is as follows: wherein, is a diagonal matrix consisting of real eigenvalues, , ; , column vector is the corresponding eigenvector of eigenvalue .

6. The similarity matrix based time series polarimetric SAR cumulative change detection method of claim 5, wherein, the maximum eigenvalue is represented as: 。 7. The similarity matrix based time series polarimetric SAR cumulative change detection method of claim 6, wherein, the difference is represented by: 。 8. The similarity matrix based time series polarimetric SAR cumulative change detection method of claim 1, wherein, The diagonalization of the similarity matrix is expressed as: The method for obtaining the cumulative change detection result specifically comprises: Constructing a Markov random field (MRF) energy function model of the difference map; Carrying out binary segmentation on the difference map based on the MRF energy function minimization of the simulated annealing global optimization algorithm to obtain the cumulative change detection result of polarimetric SAR images in a time sequence.

Images