CN107341781B

CN107341781B - SAR image correction method based on improved phase consistency feature vector base map matching

Info

Publication number: CN107341781B
Application number: CN201710594476.6A
Authority: CN
Inventors: 王峰; 向俞明; 尤红建; 刘佳音
Original assignee: Institute of Electronics of CAS
Current assignee: Zhongke Satellite Shandong Technology Group Co ltd; Aerospace Information Research Institute of CAS
Priority date: 2017-07-20
Filing date: 2017-07-20
Publication date: 2020-06-09
Anticipated expiration: 2037-07-20
Also published as: CN107341781A

Abstract

The invention provides an SAR image correction method based on improved phase consistency feature vector base map matching, which comprises the following steps: converting the vector base map into a binary grid image according to a specified resolution; geocoding the original SAR image to be corrected according to the specified resolution; determining a common area between the two images according to the rasterized vector base map and the geographic information of the SAR image after the geographic coding; extracting the linear characteristic graph of the SAR image after geocoding based on an improved phase consistency operator, and automatically matching the linear characteristic graph with a rasterized vector base graph; carrying out template matching processing on the vector base image slice and the image slice after the linear features are extracted; and screening matching point pairs by using a Ranpac method, removing mismatching points to obtain control point information of the original SAR image, and performing image correction processing by using an image space affine transformation model to obtain the SAR image with accurate geometric positioning. The correction method disclosed by the invention improves the accuracy of the matching result.

Description

SAR image correction method based on improved phase consistency feature vector base map matching

Technical Field

The disclosure relates to the technical field of remote sensing, in particular to an SAR image correction method based on improved phase consistency feature vector base map matching.

Background

Synthetic Aperture Radar (SAR) is taken as an all-weather earth observation means all day long, has obvious advantages under the conditions of disaster areas, cloud fog and the like, the number of SAR remote sensing satellites running in orbit is continuously increased in recent years, including PalSAR, Cosmo, RadarSat, TerrasAR, Gao and the like, and the obtained SAR remote sensing image data volume is larger and larger. The accurate positioning of the remote sensing image is an important premise for the quantification application of the remote sensing data, the conventional geometric correction processing of the remote sensing image needs to be carried out based on field control points, the method needs to consume a large amount of manpower and financial resources, information of an attention area cannot be obtained timely, and the timeliness is poor [ Xiao Lang, Han Chilobrachys, Liu Jia Yin ]. In recent years, an SAR image correction method based on optical control base map automatic matching is proposed, but the coverage range of an optical control image with high-precision positioning is limited, the updating of optical control data is difficult, and the automatic correction requirement for continuously updating the SAR image cannot be met.

At present, automatic geometric correction of a satellite-borne SAR remote sensing image is generally completed through automatic matching with an optical control base map, the problems of poor precision and low efficiency of an artificial geometric correction method of the remote sensing image are solved [ permit duckweed and the like ], the satellite-borne SAR image based on a GCP library is automatically and finely corrected [ J ]. surveying and mapping science, 2009,34(5): 107-; in the aspect of matching processing of an SAR image and an optical GCP slice, firstly, satellite orbit parameters are used for calculating the positions of four corners of an image to be corrected in a rough matching stage, then an affine transformation and resampling method is used for obtaining a rough matching area with uniform scale and no rotation angle, in a fine matching stage, a normalized cross-correlation coefficient is used for accurately positioning a homonymy point to obtain an accurate matching result, and the calculation formula is as follows:

wherein, (x, y) is the neighborhood point of (u, v), and E (I), E (T) are the gray level mean values of the reference image I and the template image T.

In order to realize matching processing of a raster image and a vector diagram, linear feature extraction is required, and feature extraction algorithms commonly used in this respect include a Sobel operator, a Canny operator, and the like. The Canny edge detection operator [ Canny J.Acomputational approach to edge detection [ J ]. IEEE Trans on Pattern analysis machine analysis approach 1986,8(10): 679-:

edge strength:

edge direction:

and carrying out non-extreme value suppression and edge tracking processing, and removing weak edge points to obtain a continuous edge.

For the multiplicative noise characteristics of the SAR image, the edge feature detection is implemented by a ratio detection method, such as roa (ratio of average) operator [ r.touzi, a.lopes, and p.bousquet.a static and geometry detectors for SAR images [ j.].IEEE Transaction on Geosciences and RemoteSensing,1988,26(3):764-773.]. The operator determines the gradient value of the pixel point by calculating the ratio of the pixel point to the four directions in the rain, and can better overcome the speckle noise of the SAR image. The ratio gradient operator calculates the ratio of the pixel means of two non-overlapping regions opposite on the neighborhood (fig. 1). The gray squares in the front of FIG. 1 represent the current calculated pixel points, and the ROA operator calculates the mean μ of the gray values of the pixels in the dark and light regions₁,μ₂Let r be_i＝max{μ₁/μ₂,μ₂/μ₁When r is_iThe closer to 1, the closer the mean values of the two regions are, the higher the probability that they belong to the same type of region; the larger the difference between the two regions, the more likely the point to be detected will be at the boundary between the two regions. Considering the different trend of the edge, the detection can be performed once according to the 4 directions shown in fig. 1, according to the Constant False Alarm Rate (CFAR), and the maximum value r is max { r ═ r {₁,r₂,r₃,r₄As the ROA gradient strength at that point.

Therefore, the automatic correction of the conventional satellite-borne SAR remote sensing image is mainly realized based on the automatic matching of an optical control base map, and the research of realizing the automatic correction of the satellite-borne SAR remote sensing image based on the vector base map matching is not available. With the increase of the data volume of the SAR remote sensing, how to automatically correct the SAR image by using more types of space geographic information is a problem to be solved. The SAR image linear feature extraction method is mostly realized based on the gray difference of an image domain, the edge features of the image are extracted, and a new linear feature extraction method needs to be provided aiming at the matching characteristics of the SAR image and a vector image.

The optical control image is used as a control base map for SAR image correction, and the automatic correction of the satellite-borne SAR remote sensing image is carried out through the matching processing of the optical image and the SAR image, so that the following defects exist:

1. the existing automatic correction processing of satellite-borne SAR remote sensing images is mostly realized based on an optical control base map, and the problems that the coverage range of control data is limited and the data cannot be updated in time exist. The making of the optical control base map data needs to be realized by determining ground control points through manual field surveying and mapping, a large amount of manpower and material resources need to be consumed, the data coverage range with higher confidence coefficient is limited, and a large number of areas cannot provide optical control data; furthermore, most of the optical control data used at present is historical accumulated data, and when the actual substrate changes, the control information cannot be updated in time.

2. With the continuous development of remote sensing technology, the requirements of resolution and positioning accuracy of the SAR image are continuously improved, and the image resolution unit and the positioning accuracy of the control base map based on the raster data are fixed, so that the control base map based on the raster data cannot meet the requirements of generating SAR images with different resolutions and positioning accuracies.

3. The grid data used as the control base map needs to occupy a large amount of storage space, a dedicated hard disk device is needed to store the control base map, hardware resources such as computing equipment are consumed, and the SAR automatic correction system is inconvenient to deploy in different workplaces.

4. The raster image records rich texture information, the vector image only records the positioning information of an interested target, the existing matching algorithm cannot be applied to the automatic matching processing of the SAR image and the vector image, and an algorithm suitable for the automatic matching of the SAR remote sensing image and the vector control base map needs to be designed.

5. Due to interference of multiplicative noise in the SAR image, detection effects of common line feature detection operators including Sobel operators, Canny operators and the like are poor; on the other hand, the vector underlying map (e.g., road and river vector map) marks the position of the target centerline, and thus, a general SAR edge detection operator, such as ROA, is also not applicable.

In a word, the existing SAR remote sensing image automatic correction processing method is realized based on an optical raster control image, the coverage range of the optical raster control image is limited, only fixed resolution images can be stored, and a large amount of storage space is occupied. In order to overcome the defects, an SAR image automatic correction method based on a vector base map needs to be realized, a corresponding automatic matching method is designed according to the characteristics of the SAR remote sensing image and the vector base map, and the stability and the accuracy of the automatic correction process are ensured.

Disclosure of Invention

Technical problem to be solved

In view of the technical problems, the invention provides an SAR image automatic correction method for carrying out vector base map matching based on improved phase consistency characteristics, which solves the problem of automatic correction of satellite-borne optical remote sensing images in the coverage area of traditional raster data without traditional optical control images, realizes automatic matching of SAR images and vector base maps and improves accuracy of matching results.

(II) technical scheme

According to one aspect of the disclosure, an SAR image correction method based on improved phase consistency feature vector base map matching is provided, which includes the following steps:

converting the vector base map into a binary grid image according to a specified resolution;

geocoding the original SAR image to be corrected according to the specified resolution;

determining a common area between the two images according to the rasterized vector base map and the geographic information of the SAR image after the geographic coding;

extracting a geocoded SAR image linear feature map based on an improved phase consistency feature operator for automatic matching with a rasterized vector base map;

carrying out template matching processing on the vector base image slice and the image slice after the linear features are extracted;

and screening matching point pairs by using a Ranpac method, removing mismatching points to obtain control point information of the original SAR image, and performing image correction processing by using an image space affine transformation model to obtain the SAR image with accurate geometric positioning.

In some embodiments of the present disclosure, in the step of converting the vector base map into the binarized raster image according to the specified resolution, a pixel value of a vector labeling position in the vector base map is set to 255, and a pixel of a non-vector labeling position is set to 0, so as to obtain a binarized raster image having the same resolution as that of the original SAR image to be matched.

In some embodiments of the present disclosure, a common region between the two images is determined according to the geographic information of the image to be corrected and the vector base map, and the calculation formula is as follows:

P＝F_Shp∩F_SAR

wherein, F_ShpInformation representing the geographical range of the underlying map to be vectorized, F_SARAnd representing the geographical range information of the SAR image to be corrected, wherein P is the searched common area information.

In some embodiments of the present disclosure, in the step of performing template matching processing on the vector control slice and the image slice after the linear feature is extracted, similarity of the template slices is compared by using a similarity metric criterion, so as to obtain correspondence of the slices.

In some embodiments of the disclosure, the similarity metric criteria include: normalized cross-correlation, mutual information, and phase correlation.

In some embodiments of the present disclosure, the extracting an original SAR image linear feature map to be corrected based on an improved phase consistency operator includes:

aiming at the multiplicative noise characteristics of the SAR image, improving the original local energy model by utilizing an SAR local energy model conforming to the multiplicative noise characteristics to obtain an improved phase consistency operator suitable for SAR feature extraction;

carrying out noise threshold estimation aiming at the characteristics of the SAR local energy model;

and obtaining an SAR-PC model suitable for a multiplicative noise model according to the SAR local energy model and a noise threshold value, and extracting an original SAR image linear characteristic diagram to be corrected.

In some embodiments of the present disclosure, a Gabor filter is employed as a quadrature filter bank to obtain the SAR local energy model that conforms to multiplicative noise.

In some embodiments of the present disclosure, the Gabor filter is divided into an odd symmetric component and an even symmetric component; the Gabor even symmetric component comprises two sub-windows, the Gabor odd symmetric component comprises three sub-windows, and the convolution of an image and an orthogonal filter bank in the original energy PC model is replaced by the ratio of the odd symmetric window to the even symmetric window to obtain an SAR local energy model conforming to multiplicative noise.

In some embodiments of the present disclosure, the estimating a noise threshold based on the SAR local energy model includes: noise response is eliminated based on ratio calculation of an odd symmetrical window and an even symmetrical window, noise estimation is converted into noise threshold estimation of a ratio operator, an inverse relation exists between a noise judgment threshold and a change coefficient of an image, and the noise threshold is automatically calculated through the change coefficient.

In some embodiments of the present disclosure, further comprising: and selecting a scale parameter for improving the SAR phase consistency detection operator according to the size of the actual ground object target and the image resolution information.

(III) advantageous effects

According to the technical scheme, the SAR image automatic correction method for carrying out vector base map matching based on the improved phase consistency features at least has one of the following beneficial effects:

(1) the method adopts the vector base map as the control base map, expands the types of the control base map which can be used for automatic correction processing of the satellite-borne SAR remote sensing image, and solves the problem of automatic correction of the satellite-borne SAR remote sensing image without the coverage area of the traditional optical control image.

(2) The method is suitable for an automatic matching method of a satellite-borne SAR remote sensing image and a vector base map, and solves the scale and rotation difference between images to be matched through geocoding with specified resolution; the size of the SAR phase consistency line feature detection operator is determined by combining the size of the ground object target and the image resolution, the SAR phase consistency line feature detection operator is different from the size of the target edge extracted by a common edge detection algorithm, the target center line is extracted by the method, and the accuracy of a matching result is improved corresponding to information recorded in a vector base map.

(3) The method is suitable for extracting the linear features of the SAR image conforming to the multiplicative noise model, and provides an important basis for automatic matching of the SAR image and the vector base map.

Drawings

The above and other objects, features and advantages of the present disclosure will become more apparent from the accompanying drawings. Like reference numerals refer to like parts throughout the drawings, which are not intended to be drawn to scale, emphasis instead being placed upon illustrating the principles of the disclosure.

Fig. 1 is a schematic diagram of a ROA template operator in four directions according to the prior art.

Fig. 2 is a schematic diagram of a two-dimensional Gabor filter according to an embodiment of the disclosure.

Fig. 3 is a technical flowchart of a method for automatically correcting a satellite-borne SAR remote sensing image based on vector base map matching according to an embodiment of the disclosure.

Fig. 4 is a schematic diagram of an example 1 of an automatic matching result of a satellite-borne SAR remote sensing image and a vector slice according to the present disclosure.

Fig. 5 is a schematic diagram of an example 2 of an automatic matching result of a satellite-borne SAR remote sensing image and a vector slice according to the present disclosure.

Fig. 6 is a schematic diagram of an example 3 of an automatic matching result of a satellite-borne SAR remote sensing image and a vector slice according to the present disclosure.

Fig. 7 is a schematic diagram of distribution of correction accuracy check points of the terrasaar remote sensing image.

Fig. 8a is a schematic diagram of the terrasaar raw positioning accuracy inspection.

Fig. 8b is a schematic diagram of the positioning accuracy inspection example 1 after correction according to the present disclosure.

Fig. 9a is a schematic diagram of the terrasaar raw positioning accuracy inspection.

Fig. 9b is a schematic diagram of the positioning accuracy inspection example 2 after correction according to the present disclosure.

Fig. 10a is a schematic diagram of the terrasaar raw positioning accuracy inspection.

Fig. 10b is a schematic diagram of the positioning accuracy inspection example 3 after correction according to the present disclosure.

Detailed Description

For the purpose of promoting a better understanding of the objects, aspects and advantages of the present disclosure, reference is made to the following detailed description taken in conjunction with the accompanying drawings.

It should be noted that in the drawings or description, the same drawing reference numerals are used for similar or identical parts. Implementations not depicted or described in the drawings are of a form known to those of ordinary skill in the art. Additionally, while exemplifications of parameters including particular values may be provided herein, it is to be understood that the parameters need not be exactly equal to the respective values, but may be approximated to the respective values within acceptable error margins or design constraints. Directional phrases used in the embodiments, such as "upper," "lower," "front," "rear," "left," "right," and the like, refer only to the orientation of the figure. Accordingly, the directional terminology used is intended to be in the nature of words of description rather than of limitation.

The disclosure provides an SAR image correction method based on improved phase consistency feature vector base map matching, which is an SAR image automatic correction method based on improved phase consistency feature to carry out vector base map matching. Fig. 3 is a flowchart of an automatic SAR image correction method for performing vector base map matching based on improved phase consistency characteristics according to the present disclosure. Referring to fig. 3, the method for automatically correcting an SAR image based on improved phase consistency features to perform vector base map matching includes:

and S1, converting the vector diagram (namely the vector base diagram and the control base diagram) into a binary grid image according to the specified resolution. The pixel value of the position with the vector mark is 255, and the pixel value of the position without the vector mark is 0, so that a binary grid map with the same resolution as the original SAR image to be matched can be obtained.

And S2, geocoding the original SAR image to be corrected according to the specified resolution. The specific coding mode can adopt a general correction mode according to an RPC file provided by original data or adopt an RD model correction mode based on parameters such as satellite attitude, speed and orbit information.

S3, determining a common area between the two images according to the geographic information of the rasterized vector base map in S1 and the geocoded SAR image in S2, wherein the calculation formula is as follows:

P＝F_Shp∩F_SAR

And S4, extracting the linear feature map of the SAR image after geocoding processing in S2 (namely, the improved SAR phase consistency operator line feature detection algorithm) based on the improved phase consistency operator, preferably extracting the common region part of the SAR image after geocoding processing and the vector base map, and automatically matching the common region part of the SAR image after geocoding processing and the vector base map in S1, preferably the common region part of the SAR image and the rasterized vector base map.

S41, aiming at the multiplicative noise characteristics of the SAR image, the original local energy model is improved by utilizing the SAR local energy model conforming to the multiplicative noise characteristics, an improved phase consistency operator suitable for SAR feature extraction is obtained, and the improved phase consistency operator is applied to SAR image structure feature extraction, so that the linear features of the SAR image can be extracted more stably. The conventional phase consistency operator is designed based on an additive noise model, is suitable for extracting structural information from an optical image, and has the following calculation formula:

wherein, W is frequency propagation weight quantity, theta represents orientation angle, and the value range is [0, pi ]]，A_n(x, theta) and

is the image point with the dimension n of the Log Gabor filter,the amplitude and phase at the direction theta,

is the average phase and epsilon is a small constant, only signals with energy values above the noise threshold T are included in the result.

The noise model of the image obtained by the SAR sensor is different from that of the optical image, so that the SAR local energy model conforming to the characteristic of multiplicative noise needs to be designed. Here, Gabor filters are used as the orthogonal filter bank, and the Gabor filters can be decomposed into an odd symmetric component and an even symmetric component:

from fig. 2, it can be seen that the Gabor even symmetric component includes two sub-windows, the Gabor odd symmetric component includes three sub-windows, and the average value of each sub-window in the Gabor window is recorded as

And

as shown in FIG. 2, the mean values of the Gabor even symmetric component including two sub-windows are respectively recorded as

And

the mean values of the Gabor odd symmetric component including three sub-windows are respectively recorded as

And

aiming at the characteristics of the SAR image, the convolution of the image and the orthogonal filter bank in the original phase consistency operator PC energy model is replaced by the ratio calculated by an odd symmetrical window and an even symmetrical window, as shown in the following formula:

the corresponding local energy model and amplitude values are obtained by the following method:

when the above formula is extended to a two-dimensional case, directivity needs to be considered, and six directions are often adopted for calculation in consideration of the problems of detection accuracy and calculation complexity.

S42, designing a noise threshold estimation method aiming at the local energy model characteristics of the SAR for noise threshold estimation; specifically, the SAR local energy model theoretically solves the speckle noise problem, noise response is eliminated based on ratio calculation in the SAR local energy model, and noise estimation is converted into noise judgment threshold estimation of a ratio operator. Through experiments, the noise judgment threshold value and the change coefficient of the image have an inverse relationship, so that the noise threshold value can be automatically calculated through the change coefficient, as shown in the following formula:

T＝0.6+alog(1/Cv)

Cv＝σ/μ

wherein a is a constant for adjusting the shape of the threshold curve; μ and σ denote mean and variance in the neighborhood, respectively.

S43, substituting the SAR local energy model and the noise threshold into an original Phase Consistency (PC) model, so as to obtain an SAR-PC model suitable for a multiplicative noise model:

where θ represents the angle of the Gabor window, W_θ(x, y) is the frequency propagation weight,. epsilon.is a small constant, T_sar-θ(x, y) is the noise suppression factor, and is only logged in the result if the energy value exceeds this threshold.

And S44, selecting scale parameters of the SAR phase consistency detection operator according to the size of the actual ground object target and the image resolution information. Taking a road vector diagram as an example, the vector labeling ground object target is an urban road, the road width is about ten meters, the resolution of the Terra-SAR image SpotLight mode is 2 meters, the road width is 5 pixels, and the size of the middle template with phase consistency is selected to be 5.

And S5, performing template matching processing (block template matching) on the vector control slices and the image slices after the linear features are extracted to obtain the matching point relation among the slices. And comparing the similarity of the template slices by using a similarity measurement criterion so as to obtain the corresponding relation of the slices. Similarity metric criteria may be selected herein to include normalized cross-correlation, mutual information, and phase correlation. The commonly used normalized cross-correlation method is calculated as follows:

wherein, W is the template image, E (W) is the mean value of the gray values in the template image window, and I is the image to be searched. The central pixel (i, j) of the image search window is moved continuously, and the position where the maximum value of the cross-correlation coefficient is obtained corresponds to the area with the highest similarity with the template image in the image to be searched.

S6, screening matching point pairs by using a Ranpac method, removing mismatching points to obtain control point information of the original SAR image, and performing image correction processing by using an image space affine transformation model to obtain the SAR image with accurate geometric positioning.

The method for automatically correcting the SAR image based on the improved phase consistency feature for matching the vector base map is further described below by way of example.

Firstly, an L1-grade satellite-borne SAR image product of a TerrraSAR Beijing area is selected, a public road vector diagram of the Beijing city in 2013 is used as a control base map, and then the method provided by the disclosure is adopted for automatic correction processing. The raster image and the vector base map are uniformly converted into a geocoded image with the resolution of 2 meters, and an automatic matching slice is manufactured according to the size of 800m multiplied by 800 m. The initial positioning deviation of the terrasaar image is within 50 meters, the template matching search range is set to 50/2-25 pixels, and fig. 4 to 6 show the results of automatic matching of a part of the TerraSAR image and a vector slice, wherein a in fig. 4 to 6 is a vector base map slice and a central point position, b is an original SAR image and an automatic matching position, and c is an extraction linear feature and an automatic matching position of the SAR image.

The number of control points obtained by automatic matching is 23, and automatic correction processing of the satellite-borne SAR remote sensing image is carried out based on the obtained control information. The images before and after being corrected by the method disclosed by the invention are checked for positioning accuracy, 18 uniformly distributed check points are used in total, and the distribution of the check points in the image range is shown in FIG. 7. The uncontrolled geometric positioning precision of the TerrasAR image obtained through the check point is 21.1 m, the geometric positioning precision of the image after correction processing by adopting the method is 2.5 m, and the positioning deviation is about 1.25 pixels according to the 2 m resolution.

The method comprises the steps of respectively carrying out uncontrolled correction on a Terras SAR remote sensing image and correction processing based on the method, wherein partial accuracy inspection results of the image obtained by two different processing methods are shown in FIGS. 8-10, white dotted lines mark the positions of roads marked by a vector base map, FIGS. 8a, 9a and 10a are the results of superposition display of an original SAR image and the vector map, and FIGS. 8b, 9b and 10b are the results of superposition display of the SAR image and the vector map after correction by adopting the method, and the positioning relation between the SAR image and the vector base map before and after correction is compared.

It will be appreciated by persons skilled in the art that the above implementation methods are only used for illustrating the disclosure, but not as limiting the disclosure, and that changes and modifications to the above examples are within the scope of the disclosure

The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

It is to be noted that, in the attached drawings or in the description, the implementation modes not shown or described are all the modes known by the ordinary skilled person in the field of technology, and are not described in detail. Furthermore, the above definitions of the various elements and methods are not limited to the specific structures, shapes or modes mentioned in the examples, and may be modified or substituted by one of ordinary skill in the art:

(1) although the energy model of the phase congruency operator is improved by using the neighborhood ratio method in the above embodiment to extract the central linear feature of the target of the SAR grid map, other forms of ratio methods can be used to improve the phase congruency operator, and all the methods are within the scope of the present disclosure;

(2) for example, a method for automatically correcting vector base maps of rivers, coastlines, and the like can be adopted based on a road vector map as a control base map, and the method belongs to the scope of the present disclosure.

The above-mentioned embodiments are intended to illustrate the objects, aspects and advantages of the present disclosure in further detail, and it should be understood that the above-mentioned embodiments are only illustrative of the present disclosure and are not intended to limit the present disclosure, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present disclosure should be included in the scope of the present disclosure.

Claims

1. An SAR image correction method based on improved phase consistency feature vector base map matching comprises the following steps:

extracting the linear characteristic graph of the SAR image after geocoding based on an improved phase consistency operator, and automatically matching the linear characteristic graph with a rasterized vector base graph;

screening matching point pairs by using a Ranpac method, removing mismatching points to obtain control point information of an original SAR image, and performing image correction processing by using an image space affine transformation model to obtain an SAR image with accurate geometric positioning;

the method for extracting the linear characteristic diagram of the original SAR image to be corrected based on the improved phase consistency operator comprises the following steps:

obtaining an SAR-PC model suitable for a multiplicative noise model according to the SAR local energy model and a noise threshold value, and extracting an original SAR image linear characteristic diagram to be corrected;

and selecting the scale parameter of the improved phase consistency operator according to the size of the actual ground object target and the image resolution information.

2. The SAR image correction method according to claim 1, wherein in the step of converting the vector base map into the binarized raster image according to the specified resolution, a pixel value of a vector labeling position in the vector base map is 255, and a pixel of a non-vector labeling position is 0, so as to obtain a binarized raster image having the same resolution as that of the original SAR image to be matched.

3. The SAR image correction method of claim 1, wherein a common region between two images is determined according to geographic information of the image to be corrected and the vector base map, and a calculation formula is as follows:

P＝F_Shp∩F_SAR

4. The SAR image correction method according to claim 1, wherein in the step of performing template matching processing on the vector control slice and the image slice after the linear feature extraction, similarity of the template slices is compared by using a similarity measurement criterion, so as to obtain correspondence of the slices.

5. The SAR image correction method of claim 4, wherein the similarity metric criteria comprises: normalized cross-correlation, mutual information, and phase correlation.

6. The SAR image correction method of claim 5,

and obtaining the SAR local energy model conforming to the multiplicative noise by adopting a Gabor filter as an orthogonal filter bank.

7. The SAR image correction method of claim 6,

the Gabor filter is divided into an odd symmetrical component and an even symmetrical component; the Gabor even symmetric component comprises two sub-windows, the Gabor odd symmetric component comprises three sub-windows, and the convolution of an image and an orthogonal filter bank in the original energy PC model is replaced by the ratio of the odd symmetric window to the even symmetric window to obtain an SAR local energy model conforming to multiplicative noise.

8. The SAR image correction method of claim 5,

the estimating a noise threshold value based on the SAR local energy model comprises the following steps: noise response is eliminated based on ratio calculation of an odd symmetrical window and an even symmetrical window, noise estimation is converted into noise threshold estimation of a ratio operator, an inverse relation exists between a noise judgment threshold and a change coefficient of an image, and the noise threshold is automatically calculated through the change coefficient.