Signal processing system, signal processing method, and signal processing program

Abstract

The present invention provides a signal processing system, a signal processing method, and a signal processing program capable of selectively extracting pixel groups from a complex image. [Solution] The system includes an assignment means for assigning each pixel to a cluster such that a complex scale indicating the degree of match between a complex number vector obtained from corresponding pixels of multiple complex images and the probability parameter of the cluster is maximized, and a parameter calculation means for calculating a probability parameter for each cluster such that the complex scale is maximized for all assigned pixels, wherein the assignment means and the parameter calculation means operate alternately.

Description

[Technical Field]

[0001] This disclosure relates to a signal processing system, a signal processing method, and a signal processing program. [Background technology]

[0002] As a technology related to signal processing, for example, Patent Document 1 describes a technique for extracting statistically homogeneous pixels from SAR images obtained using synthetic aperture radar (SAR) technology.

[0003] The pixel identification device described in Patent Document 1 considers that a pixel of interest and its neighbors follow the same probability distribution if the maximum absolute value of the difference between the cumulative density function of the pixel of interest and the cumulative density function of neighboring pixels is smaller than a predetermined threshold. In other words, the pixel identification device considers that the pixel of interest and its neighbors are generated with the same probability density function and determines that the neighboring pixels are statistically homogeneous with the pixel of interest. [Prior art documents] [Patent Documents]

[0004] [Patent Document 1] International Publication No. 2010 / 112426 [Overview of the Initiative] [Problems that the invention aims to solve]

[0005] SAR images are complex images that contain information on the reflection intensity and phase of the irradiated microwaves for each pixel. However, the invention described in Patent Document 1 does not consider the similarity of phases between pixels, even though statistically homogeneous pixels are extracted. As a result, the extracted pixel group may contain pixels with dissimilar phases. Consequently, using these pixel groups for phase analysis may lead to inaccurate analysis results.

[0006] This disclosure has been made in view of these issues. One of the objectives of this disclosure is to provide a signal processing system, a signal processing method, and a signal processing program that can suitably extract pixel groups from a complex image. [Means for solving the problem]

[0007] The signal processing system according to this disclosure includes an assignment means for assigning each pixel to a cluster such that a complex scale indicating the degree of match between a complex number vector obtained from corresponding pixels of a plurality of complex images and the probability parameter of the cluster is maximized, and a parameter calculation means for calculating a probability parameter for each cluster such that the complex scale is maximized for all assigned pixels, wherein the assignment means and the parameter calculation means operate alternately.

[0008] The signal processing method according to this disclosure involves a computer assigning each pixel to a cluster such that the complex scale indicating the degree of match between the complex number vector obtained from corresponding pixels of multiple complex images and the probability parameter of the cluster is maximized, and for each cluster, calculating the probability parameter such that the complex scale is maximized for all assigned pixels, and performing the assignment and calculation alternately.

[0009] The signal processing program according to this disclosure causes a computer to perform an assignment process, which assigns each pixel to a cluster such that the complex scale indicating the degree of match between the complex number vector obtained from corresponding pixels of multiple complex images and the probability parameter of the cluster is maximized, and a parameter calculation process, which calculates the probability parameter for each cluster such that the complex scale is maximized for all assigned pixels, and the assignment process and the parameter calculation process are executed alternately. [Effects of the Invention]

[0010] According to this disclosure, a group of pixels can be suitably extracted from a complex image. [Brief explanation of the drawing]

[0011] [Figure 1] It is a block diagram illustrating a signal processing device. [Figure 2] It is an explanatory diagram showing an example of a SAR image. [Figure 3] It is a flowchart showing an operation example of the signal processing device. [Figure 4] It is a flowchart showing another operation example of the signal processing device. [Figure 5] It is a block diagram illustrating a signal processing device. [Figure 6] It is a block diagram illustrating a signal processing system. [Figure 7] It is a flowchart illustrating the operation of the signal processing system. [Figure 8] It is a block diagram illustrating a signal processing system. [Figure 9] It is a block diagram illustrating the operation of the signal processing system. [Figure 10] It is a block diagram illustrating the hardware configuration of a computer. [Figure 11] It is a block diagram illustrating the main part of the signal processing system.

Embodiments for Carrying Out the Invention

[0012] Synthetic Aperture Radar (SAR) technology is a technology for obtaining a SAR image equivalent to an image by an antenna with a large aperture while a flying object such as a satellite or an aircraft moves and the radar mounted on the flying object transmits and receives electromagnetic waves. The synthetic aperture radar is used, for example, for analyzing surface displacement and the like by processing reflected waves from the surface. Note that the surface includes not only the ground but also the surface (upper surface) where low structures such as buildings exist.

[0013] An image taken by a flying object such as a satellite is called a radar image. The SAR image is an example of a radar image. Hereinafter, it is assumed that the flying object that transmits and receives electromagnetic waves is a satellite, but the flying object is not limited to a satellite.

[0014] One example of image analysis is interferometry, which investigates displacement, elevation, etc., based on the phase difference between multiple SAR images. Another example of image analysis is change detection, which detects changes or anomalies on the ground based on changes in intensity. These are just a few examples of image analysis, and the field of image analysis is diverse.

[0015] SAR images contain natural and artificial objects of various properties. Therefore, the pixel values ​​of multiple pixels in a SAR image may have different properties depending on the object that the pixel is depicting. In particular, the probabilistic variation of pixel values, i.e., noise, is known to have different characteristics depending on the type of subject. Therefore, there is a problem in that the results of image analysis may be inaccurate if the characteristics dependent on the type of subject are not taken into consideration.

[0016] To solve such problems, extracting statistically homogeneous pixels is useful. For example, Patent Document 1 describes a technique for extracting statistically homogeneous pixels.

[0017] However, the pixel identification device described in Patent Document 1 does not consider the phase similarity between pixels when extracting pixels. Therefore, the extracted pixel group may contain pixels with dissimilar phases. As a result, if these pixel groups are used for phase analysis, the analysis results may be inaccurate. Furthermore, the pixel identification device described in Patent Document 1 requires an enormous amount of computation time because it compares each of the many pixels of interest with all the pixels in the window. In other words, it takes a long time to obtain a statistically homogeneous pixel group. This disclosure has been made in view of these problems.

[0018] Embodiments of this disclosure will be described below with reference to the drawings. In each drawing, the same or corresponding elements are denoted by the same reference numerals, and redundant explanations are omitted as necessary for clarity. Unless otherwise specified, predetermined values ​​such as set values ​​and thresholds are stored in advance in a storage device accessible from the device that uses those values. Unless otherwise specified, the storage unit is composed of one or any number of storage devices.

[0019] In the embodiments described below, SAR images are used as an example of radar images obtained using electromagnetic waves. However, radar images are not limited to SAR images. For example, a radar image may be an optical image. Furthermore, a SAR image is a complex image in which each pixel has information about the reflection intensity and phase of the irradiated microwave as a pixel value. Hereafter, a SAR image will also be referred to as a complex image.

[0020] Embodiment 1. Figure 1 is a block diagram illustrating a signal processing device. The signal processing device 100 in this embodiment includes a pixel allocation unit 110, a parameter calculation unit 120, an allocation information storage unit 130, a parameter information storage unit 140, and an output unit 150. The signal processing device 100 can receive SAR images from a SAR image storage unit 200. The SAR image storage unit 200 may be included in the signal processing device 100 or in a device other than the signal processing device 100.

[0021] The signal processing device 100 is a signal processing device (similar pixel extraction device) in the signal processing system that extracts similar pixels in a SAR image. In other embodiments described later, the signal processing device also constitutes a similar pixel extraction device that extracts similar pixels in a SAR image.

[0022] The pixel assignment unit 110 has the function of assigning each pixel of the SAR image to a cluster. For example, the pixel assignment unit 110 assigns each pixel to one of several clusters such that the complex scale indicating the degree of match between the complex vector obtained from corresponding pixels of multiple complex images and the probability parameter of the cluster is maximized.

[0023] The parameter calculation unit 120 has a function to calculate the probability parameters of the clusters. For example, for each cluster, the parameter calculation unit 120 calculates the probability parameters such that the complex scale is highest for the entire set of pixels to which the parameter is assigned.

[0024] The allocation information storage unit 130 stores allocation information indicating pixels assigned to clusters by the pixel allocation unit 110. The allocation information includes, for example, information indicating the correspondence between clusters and the pixels assigned to those clusters. Hereinafter, pixels assigned to clusters will also be referred to as pixels belonging to clusters.

[0025] The parameter information storage unit 140 stores parameter information that shows the probability parameters for each cluster calculated by the parameter calculation unit 120.

[0026] The output unit 150 has a function to output assignment information and parameter information. For example, the output unit 150 outputs and stores the assignment information and parameter information in the signal processing device 100 or the storage unit (not shown) of an external device. Alternatively, the output unit 150 outputs and displays the assignment information and parameter information on a display device (not shown), such as a display device. The output unit 150 may output only one of the assignment information and parameter information, or both.

[0027] The SAR image storage unit 200 stores S images (for example, S = 10 to 30) of the same area. In other words, the SAR image storage unit 200 stores S images that show the same analysis area. Hereinafter, the SAR images (SAR image group) stored in the SAR image storage unit 200 will also be referred to as input SAR images or input SAR image group. The SAR images are aligned so that pixels at the same position in each SAR image correspond to the same point or feature.

[0028] In Figure 1, the arrows simply illustrate the flow of signals (data), but there is no intention to exclude bidirectional communication. This is also true for other block diagrams.

[0029] Each of the S SAR images is a radar image recording the same area, but acquired at different times or trajectories. The SAR images stored in the SAR image storage unit 200 may be images acquired at different times but on the same trajectory. Furthermore, multiple SAR images may be images acquired at the same time but on different trajectories. Moreover, multiple SAR images may be images with different acquisition times and acquisition trajectories.

[0030] The imaging conditions (acquisition time, incident angle, bandwidth, etc.) are not limited to the conditions under which the image was actually taken, but may be artificially synthesized. For example, in an imaging method called polarization SAR, by controlling conditions such as the electric field direction of the irradiated electromagnetic wave and the sensitivity and phase delay of the antenna receiving the electromagnetic wave (polarization imaging conditions), it is possible to obtain characteristics that depend on the electric field direction of the electromagnetic wave. In this imaging method, by synthesizing images taken under two to three different polarization conditions, it is possible to reproduce an image taken under any desired polarization condition.

[0031] Furthermore, even with regular SAR (non-polarized SAR), it is possible to extract images of only a specific frequency band by performing a frequency domain transformation, such as a Fourier transform, on the captured SAR image and then applying a filter to extract a portion of the frequency band. This filtering process may involve randomly selecting a portion of the frequency band or randomly excluding a portion of the frequency band.

[0032] By using the methods described above, it is possible to construct a set of SAR images captured in multiple different frequency bands. In other words, it is possible to generate images under imaging conditions that are not physically used and construct a set of SAR images.

[0033] Figure 2 is an explanatory diagram showing an example of SAR images. Figure 2 shows S SAR images from Image 1 to Image S. Each SAR image is aligned so that pixels at the same location correspond to the same point or feature.

[0034] In this disclosure, a complex vector is a feature vector obtained from corresponding pixels in multiple SAR images. Corresponding pixels are pixels at the same location in each SAR image, and are pixels of the same point or feature.

[0035] A complex vector has as many elements as there are inputs, corresponding to each pixel in a given set of aligned SAR images. For example, the complex vector x corresponds to each pixel with pixel number p. p (→) is represented as shown in equation (1) below. Note that (→) represents a vector symbol.

[0036]

number

[0037] The complex scale in this disclosure is a measure of the degree to which a pixel belongs to a certain cluster. An example using equation (2) below as the complex scale will be explained below.

[0038]

number

[0039] In equation (2), c is the cluster number used to identify the cluster. Γ c x is the covariance parameter corresponding to cluster c. p is the pixel number used to identify the pixel. S is the total number of SAR images. p (→) is a complex vector obtained from each pixel corresponding to pixel number p.

[0040] Equation (2) is a probability density function based on a 0-mean multivariate complex Gaussian distribution. The signal processing device 100 uses Γ as a complex measure to determine the degree to which the pixel with pixel number p belongs to cluster c. c We use a multivariate complex Gaussian model with covariance as the probability parameter.

[0041] Note that the complex scales that the signal processing device 100 can use are not limited to those shown in equation (2). The signal processing device 100 can use various complex scales. However, this complex scale must relate to both the phase and the intensity (i.e., the absolute value of the complex number) as a complex number. Also, this complex scale is x p It is desirable that the integral of (→) is constant regardless of the value of the probability parameter.

[0042] For example, the signal processing device 100 may use equation (3) below as the complex scale.

[0043]

number

[0044] In equation (3), A C Some elements are always constrained to 0. Elements constrained to 0 may be associated with, for example, pairs of dates that are far apart or pairs of orbits that are far apart.

[0045] In equation (3), the probability parameter is the inverse of the variance-covariance matrix, and is constrained to being partially zero. That is, equation (3) imposes a sparse constraint on the precision matrix, which is the inverse of the variance-covariance matrix, such that many of its elements are zero. By using the complex scale shown in equation (3), the signal processing device 100 can obtain the effect of reducing the computational complexity and stabilizing the computation.

[0046] Furthermore, the signal processing device 100 may use equation (4) below as the complex scale.

[0047]

number

[0048] Equation (4) is the probability density function of the multivariate complex t-distribution. In equation (4), ν is a fixed value, not a probability parameter, and is used to control robustness to outliers. The signal processing device 100 can control its robustness to outliers by using the complex scale shown in equation (4).

[0049] Furthermore, not limited to the examples above, for example, the signal processing device 100 may have x p Alternatively, the phase of (→) may be processed using a von Mises distribution and the absolute value using a Rice distribution.

[0050] Next, the operation of the signal processing device 100 will be described. Figure 3 is a flowchart showing an example of the operation of the signal processing device 100.

[0051] The signal processing device 100 takes multiple aligned SAR images as input and performs an initialization process. That is, the signal processing device 100 divides the SAR images into multiple initial clusters. The method of division is arbitrary. As an example, the signal processing device 100 creates multiple initial clusters such that the area of ​​each cluster is the same. Then, the signal processing device 100 sets a probability parameter for each cluster (step S101). The process in step S101 may be performed by the parameter calculation unit 120 of the signal processing device 100.

[0052] Next, the pixel allocation unit 110 assigns each pixel of the SAR image to one of the clusters such that the complex scale is highest (step S102). In other words, the pixel allocation unit 110 creates multiple clusters. The pixel allocation unit 110 also stores allocation information indicating the pixels assigned to the clusters in the allocation information storage unit 130 based on the allocation results.

[0053] Next, the parameter calculation unit 120 calculates probability parameters for each cluster such that the complex scale is highest for all pixels assigned to it (step S103). The parameter calculation unit 120 also stores parameter information indicating the probability parameters for each cluster in the parameter information storage unit 140 based on the calculation results.

[0054] Next, the signal processing device 100 determines whether a predetermined termination condition is met (step S104). If the predetermined termination condition is not met, the signal processing device 100 returns to the process in step S102. If the predetermined termination condition is met, the output unit 150 outputs the assignment information stored in the assignment information storage unit 130 and the parameter information stored in the parameter information storage unit 140 (step S105). After that, the signal processing device 100 terminates its processing.

[0055] The predetermined termination condition is met, for example, when the processes in steps S102 to S103 are executed a predetermined number of times. The predetermined termination condition may also be met when the shape of each cluster created in the process of step S102 is no longer significantly different from the shape of each cluster created in the process executed in the previous time (in the previous processing loop). Alternatively, the predetermined termination condition may be met when the probability parameter of each cluster calculated in the process of step S103 is no longer significantly different from the probability parameter calculated in the previous processing loop.

[0056] Each cluster determined when the predetermined termination conditions are met corresponds to a cluster that satisfies the criteria for determining homogeneity. In other words, this cluster is a group of pixels with variability that follows the same probability distribution.

[0057] In this way, the signal processing device 100 can divide the pixels of a complex image into clusters with similar phase, intensity, and noise distribution by repeatedly executing the processes in steps S102 to S103. This is equivalent to extracting pixel groups from a complex image.

[0058] Next, other examples of operation of the signal processing device 100 will be described with reference to Figure 4. Figure 4 is a flowchart showing other examples of operation of the signal processing device 100.

[0059] The signal processing device 100 receives multiple aligned SAR images as input and, as an initialization process, assigns each pixel to one of the clusters (step S111). That is, the signal processing device 100 divides the SAR image into multiple initial clusters. The method of division is arbitrary. As an example, the signal processing device 100 creates multiple initial clusters such that the area of ​​each cluster is the same. The process in step S111 may be performed by the pixel assignment unit 110 of the signal processing device 100.

[0060] Next, the parameter calculation unit 120 calculates probability parameters for each cluster such that the complex scale is highest for all pixels assigned to it (step S112). The parameter calculation unit 120 also stores parameter information indicating the probability parameters for each cluster in the parameter information storage unit 140 based on the calculation results.

[0061] Next, the pixel allocation unit 110 assigns each pixel of the SAR image to one of the clusters such that the complex scale is highest (step S113). In other words, the pixel allocation unit 110 creates multiple clusters. The pixel allocation unit 110 also stores allocation information indicating the pixels assigned to the clusters in the allocation information storage unit 130 based on the allocation results.

[0062] Next, the signal processing device 100 determines whether a predetermined termination condition is met (step S114). The signal processing device 100 can apply a predetermined termination condition similar to the operation example shown in Figure 3, for example. If the predetermined termination condition is not met, the signal processing device 100 returns to the process in step S112. If the predetermined termination condition is met, the output unit 150 outputs the assignment information stored in the assignment information storage unit 130 and the parameter information stored in the parameter information storage unit 140 (step S115). After that, the signal processing device 100 terminates its processing.

[0063] The operational examples shown in Figures 3 and 4 do not limit the operation of the signal processing device 100 of this disclosure.

[0064] For example, when the pixel assignment unit 110 searches for which cluster to assign a pixel to, it may limit the clusters to be searched based on the pixel's position information. For example, if the pixel position that is the centroid of each cluster is identified, the pixel assignment unit 110 may search only for clusters whose centroid is within a certain distance from the pixel position for each pixel. By limiting the clusters to be searched, the number of calculations related to the complex scale can be reduced, and the operation of the pixel assignment unit 110 can be sped up. Alternatively, for example, each pixel may have a predetermined list of potential cluster candidates to which it may be assigned, depending on its position. In this case, the cluster candidates corresponding to each pixel may not be changed during the optimization calculation process in steps S102-S103 and S112-S113.

[0065] Next, the effects of this embodiment will be described. In this embodiment, the pixel assignment unit 110 assigns each pixel of the complex image to a cluster such that the complex scale indicating the degree of match between the complex number vector and the cluster's probability parameter is maximized. The parameter calculation unit 120 calculates probability parameters for each cluster such that the complex scale is maximized for all assigned pixels. The pixel assignment unit 110 and the parameter calculation unit 120 then operate alternately until a predetermined termination condition is met. Each cluster determined when the predetermined termination condition is met corresponds to a cluster that satisfies the criteria for determining homogeneity. In other words, the pixels of the complex image are divided into clusters with similar phase, intensity, and noise distribution. With this configuration, the signal processing device 100 can suitably extract groups of pixels with variations following the same probability distribution from the complex image.

[0066] The signal processing device 100 uses the complex scale described above to perform processing using a probability distribution based on pixel values ​​expressed as complex numbers. As a result, clusters containing pixels with uniform phase mean and phase variance are created. In this way, the signal processing device 100 can improve the noise immunity of the clusters.

[0067] Furthermore, when extracting pixel groups, the signal processing device 100 does not need to perform a process of comparing each of the many pixels of interest with all the pixels in the window, as is done with the pixel identification device described in Patent Document 1. Therefore, the signal processing device 100 can quickly extract pixel groups that have variability according to the same probability distribution from a complex image.

[0068] The pixel group extracted by the signal processing device 100 can be utilized in various ways. For example, this pixel group is used in the field of interferometry, which investigates displacement, elevation, etc., based on the phase difference between multiple SAR images. Specifically, it can be used as the pixel group to be averaged when calculating an average phase that reduces the amount and influence of noise included in the phase difference. By utilizing it in this way, high-precision displacement analysis becomes possible even in noisy areas.

[0069] Furthermore, this pixel cluster is also used in the field of change detection, which detects changes and anomalies on the ground based on changes in intensity. Specifically, it can be used as multiple samples to estimate the distribution of noise levels, etc., for the purpose of calculating the amount of noise in pixel values ​​that can be observed during normal, unchanging conditions. By utilizing it in this way, highly reliable change detection becomes possible.

[0070] Furthermore, by grouping multiple pixels with similar distributions, they can be considered as a cluster of pixels (Super pixels) that are highly likely to depict the same object. By performing analysis based on the differences between these clusters of pixels, faster segmentation and other analyses become possible.

[0071] Embodiment 2. FIG. 5 is a block diagram showing a signal processing apparatus 101 according to another embodiment. The signal processing apparatus 101 includes a pixel assignment unit 110, a parameter calculation unit 121, an assignment information storage unit 130, a parameter information storage unit 140, and an output unit 150. The signal processing apparatus 101 can input a SAR image from a SAR image storage unit 200. Note that the SAR image storage unit 200 and the prior distribution storage unit 210 may be included in the signal processing apparatus 100, or may be included in a device different from the signal processing apparatus 100. The functions of the components other than the parameter calculation unit 121 in the signal processing apparatus 101 are the same as those of the components in the signal processing apparatus 100 shown in FIG. 1. Hereinafter, mainly the differences from the first embodiment will be described, and the description of the same parts will be omitted.

[0072] The parameter calculation unit 121 can input data indicating a prior distribution for the probability parameters of the clusters from the prior distribution storage unit 210. The parameter calculation unit 121 calculates probability parameters using this prior distribution. The prior distribution is represented by, for example, the following equation (5). Note that the following equation (5) is an example of a prior distribution. The parameter calculation unit 121 is not limited to the form shown in equation (5), and can use various other forms of prior distributions.

[0073] [Number]

[0074] The prior distribution of equation (5) is a distribution assumed in advance for the probability parameter Γ of cluster c c and indicates a prior assumption of what values Γ c can take. The prior distribution of equation (5) is obtained by substituting Γ c into x in the equation defined by the Complex Inverse Wishart Distribution and grouping the parts that do not depend on Γ c into Z, and is the conjugate prior distribution of equation (2). Therefore, by using the prior distribution of equation (5) together with equation (2), an optimal Γ cThis simplifies the calculation of the result. Note that the Complex Inverse Wishart Distribution is the conjugate prior distribution for equation (2), and the Complex Wishart Distribution is the conjugate prior distribution for equation (3), and using both simplifies the calculation. However, the form of the prior distribution is not limited to the conjugate prior distribution. The parameter calculation unit 121 may use prior distributions other than the conjugate prior distribution.

[0075] The method for determining the parameters of the prior distribution is arbitrary. For example, the parameters of the prior distribution may be determined according to the user's input. Alternatively, a computer such as the signal processing device 101 may use the pixels of the entire input SAR image to generate Γ c You may determine the equivalent and reflect the result in the parameters of the prior distribution.

[0076] Covariance matrix Γ based on actual data c When calculating this, the calculation results can become unstable, especially when there is little data or when there is large variation between samples. Therefore, the signal processing device 101 mitigates this instability by using a prior distribution.

[0077] In this embodiment, the parameter calculation unit 121 calculates the probability parameter using a prior distribution for the probability parameter. With this configuration, in addition to the effects of the first embodiment, the effect of being able to stably calculate the probability parameter can be obtained. [Examples]

[0078] The following describes a specific example of a signal processing system to which the signal processing device (similar pixel extraction device) realized in the above embodiment is applied.

[0079] Example 1. Figure 6 is a block diagram showing the signal processing system of the first embodiment. The signal processing system 400 of the first embodiment comprises a signal processing device 100, a displacement analysis unit 410, and a display unit 420. A signal processing device 101 may be used instead of the signal processing device 100. The signal processing device 101 may also be used in Embodiment 2 described below. Furthermore, the signal processing system 400 may be implemented by a single device.

[0080] The displacement analysis unit 410 has a function to analyze a cluster using the cluster's probability parameters. For example, the displacement analysis unit 410 receives assignment information and parameter information output from the output unit 150 of the signal processing device 100. The displacement analysis unit 410 uses the cluster's probability parameters to analyze the displacement, elevation, etc., for each cluster.

[0081] The display unit 420 is implemented by a display device such as a display device. The cluster analysis results from the displacement analysis unit 410 are displayed at the positions of pixels belonging to the cluster. For example, the displacement analysis unit 410 controls the display unit 420 to display the analysis results for each cluster at the positions of pixels assigned to the cluster.

[0082] Next, the signal processing system of this embodiment will be described. Figure 7 is a flowchart illustrating the operation of the signal processing system 400. Note that the operation of the signal processing device 100 illustrated in Figures 3 and 4 is omitted in the flowchart shown in Figure 7.

[0083] The displacement analysis unit 410 receives the assignment information and parameter information output from the output unit 150 of the signal processing device 100. The displacement analysis unit 410 uses the cluster probability parameters to analyze the displacement, elevation, etc., for each cluster (step S401).

[0084] Next, the display unit 420 displays the analysis results for each cluster performed by the displacement analysis unit 410 at the pixel positions assigned to that cluster (step S402). For example, the displacement analysis unit 410 controls the display unit 420 to display the analysis results for each cluster at the pixel positions assigned to that cluster.

[0085] Example 2. Figure 8 is a block diagram showing the signal processing system of the second embodiment. The signal processing system 401 of the second embodiment includes a signal processing device 100, a change detection unit 430, and a display unit 421. Note that the signal processing system 401 may be implemented by a single device.

[0086] The change detection unit 430 has a function to detect changes in a cluster using the cluster's probability parameters. For example, the change detection unit 430 receives assignment information and parameter information output from the output unit 150 of the signal processing device 100. The change detection unit 430 performs change detection for each cluster using the cluster's probability parameters.

[0087] The signal processing system 401 can receive a new SAR image related to a past SAR image from which a group of pixels was extracted (i.e., a cluster was created) by the signal processing device 100. In this case, the signal processing system 401 may not input the new SAR image to the signal processing device 100, but instead input it to the change detection unit 430. Alternatively, the change detection unit 430 may perform change detection for each cluster between the past SAR image and the new SAR image using assignment information and parameter information based on the past SAR image.

[0088] The display unit 421 is implemented by a display device such as a display device. The change detection results of the cluster by the change detection unit 430 are displayed at the positions of pixels belonging to the cluster. For example, the change detection unit 430 controls the display unit 421 to display the change detection results for each cluster at the positions of pixels assigned to the cluster.

[0089] Next, the signal processing system of this embodiment will be described. Figure 9 is a flowchart illustrating the operation of the signal processing system 401. Note that the operation of the signal processing device 100 illustrated in Figures 3 and 4 is omitted in the flowchart shown in Figure 9.

[0090] The signal processing system 401 receives a new SAR image (step S411) that is related to a past SAR image from which a group of pixels was extracted (i.e., a cluster was created) by the signal processing device 100.

[0091] Next, the change detection unit 430 applies the cluster assignment results from past SAR images to the new SAR image (step S412).

[0092] Next, the change detection unit 430 compares the new SAR image with past SAR images for each cluster to detect changes (step S413).

[0093] Next, the display unit 421 displays the detection result for each cluster by the change detection unit 430 at the pixel position assigned to that cluster (step S414). For example, the change detection unit 430 controls the display unit 421 to display the detection result for each cluster at the pixel position assigned to that cluster.

[0094] The operational examples shown in Figures 7 and 9 are not intended to limit the operation of the signal processing system of this disclosure.

[0095] As described above, the signal processing devices 100 and 101 of the above embodiment can suitably and quickly extract multiple pixels with variations following the same probability distribution from a complex image. That is, multiple pixels with variations following the same probability distribution are grouped together as a small number of clusters. Therefore, in the above embodiment, by replacing pixel-by-pixel analysis with cluster-by-cluster analysis, analyses such as displacement analysis, coverage classification, and anomaly detection can be performed suitably and quickly.

[0096] In the above embodiments and examples, SAR images were used as the example of images, but the above embodiments and examples can be applied to any image or point cloud that has different distribution characteristics for each pixel.

[0097] Each component in the above embodiments and examples can be configured with one piece of hardware, but can also be configured with one piece of software. Furthermore, each component can be configured with multiple pieces of hardware, or with multiple pieces of software. In addition, some parts of each component can be configured with hardware, and other parts with software.

[0098] Each function (each process) in the above embodiment can be implemented by a computer having a processor, memory, etc. For example, a program for implementing the method (process) in the above embodiment may be stored in a storage device (storage medium), and each function may be implemented by executing the program stored in the storage device with a processor.

[0099] Figure 10 is a block diagram illustrating the hardware configuration of computer 1000. Computer 1000 is any computer. For example, computer 1000 is a stationary computer such as a personal computer or a server machine. Alternatively, computer 1000 is a portable computer such as a smartphone or a tablet device. Computer 1000 may be a dedicated computer designed to implement a signal processing device or signal processing system, or it may be a general-purpose computer.

[0100] Computer 1000 has a processor 1001, a storage device 1002, memory 1003, a bus 1004, an input / output interface 1005, and a network interface 1006.

[0101] Processor 1001 is a variety of processing units, including CPUs (Central Processing Units), GPUs (Graphics Processing Units), FPGAs (Field-Programmable Gate Arrays), and DSPs (Digital Signal Processors).

[0102] The storage device 1002 is, for example, a non-transitory computer-readable medium. Non-transitory computer-readable media include various types of tangible storage media. Specific examples of non-transitory computer-readable media include semiconductor memory (e.g., mask ROM, PROM (Programmable ROM), EPROM (Erasable PROM), flash ROM).

[0103] Memory 1003 is a main memory system implemented using RAM (Random Access Memory) or similar technologies. Memory 1003 temporarily stores data when the processor 1001 executes processing.

[0104] Bus 1004 is a data transmission path for the processor 1001, memory 1003, storage device 1002, input / output interface 1005, and network interface 1006 to send and receive data to and from each other. However, the method of connecting the processor 1001 and the others to each other is not limited to bus connection.

[0105] The input / output interface 1005 is an interface for connecting the computer 1000 with input / output devices. For example, input devices such as keyboards and output devices such as display devices are connected to the input / output interface 1005.

[0106] The network interface 1006 is an interface for connecting computer 1000 to a network. This network may be a LAN (Local Area Network) or a WAN (Wide Area Network).

[0107] The storage device 1002 stores a program that implements each of the functional components in the above-described embodiments and examples. The processor 1001 reads this program into the memory 1003 and executes it to implement each of the functional components in the above-described embodiments and examples.

[0108] The signal processing device and signal processing system may be implemented using one computer 1000 or multiple computers 1000. In the latter case, the configuration of each computer 1000 does not need to be identical and can be different.

[0109] Each functional component in the above embodiments and examples may be implemented by a combination of the hardware and software described above, or by hardware (for example, a hardwired electronic circuit).

[0110] Next, an overview of the present disclosure will be described. Figure 11 is a block diagram illustrating the main parts of a signal processing system. The signal processing system 10 shown in Figure 11 (which can be implemented by, for example, a signal processing device 100 or 101, a signal processing system 400 or 401) includes an assignment means 11 (implemented by a pixel assignment unit 110 in an embodiment) that assigns each pixel to a cluster such that the complex scale indicating the degree of match between the complex number vector obtained from corresponding pixels of a plurality of complex images and the probability parameters of the cluster is maximized, and a parameter calculation means 12 (implemented by a parameter calculation unit 120 or 121 in an embodiment) that calculates the probability parameters for each cluster such that the complex scale is maximized for all assigned pixels. The assignment means 11 and the parameter calculation means 12 operate alternately. With this configuration, the signal processing system 10 can suitably extract groups of pixels from a complex image that have variability according to the same probability distribution.

[0111] The signal processing system 10 uses the complex scale described above to perform processing using a probability distribution based on pixel values ​​expressed as complex numbers. As a result, clusters containing pixels with uniform phase mean and phase variance are created. In this way, the signal processing system 10 can improve the noise immunity of the clusters.

[0112] The signal processing system 10 does not need to perform a comparison process with all pixels in the window for each of the many pixels of interest when extracting a group of pixels, as is done with the pixel identification device described in Patent Document 1. Therefore, the signal processing system 10 can quickly extract a group of pixels from a complex image that have variability according to the same probability distribution.

[0113] Although the present disclosure has been described above with reference to embodiments, the present disclosure is not limited to the embodiments described above. Various modifications to the structure and details of the present disclosure are possible, as can be understood by those skilled in the art within the scope of the present disclosure. Each embodiment can be combined with other embodiments as appropriate.

[0114] Each drawing is merely illustrative to illustrate one or more embodiments. Each drawing may be associated with one or more other embodiments rather than with only one specific embodiment. As those skilled in the art will understand, various features or steps described with reference to any one drawing can be combined with features or steps shown in one or more other drawings, for example, to create embodiments not explicitly shown or described. Not all features or steps shown in any one drawing to illustrate an exemplary embodiment are necessarily required, and some features or steps may be omitted. The order of steps shown in any of the drawings may be changed as appropriate.

[0115] Some or all of the above embodiments may also be described as follows, but are not limited to the following:

[0116] (Note 1) An assignment means for assigning each pixel to a cluster such that a complex measure indicating the degree of match between a complex number vector obtained from corresponding pixels of multiple complex images and the probability parameter of the cluster is maximized, For each cluster, the system includes a parameter calculation means for calculating the probability parameter such that the complex scale is highest for the entire set of pixels assigned to it. The assignment means and the parameter calculation means operate alternately. A signal processing system characterized by the following:

[0117] (Note 2) The aforementioned complex scale is a 0-mean multivariate complex Gaussian distribution, The aforementioned probability parameter is the variance-covariance matrix of the multivariate complex Gaussian distribution or the inverse matrix of said variance-covariance matrix. The signal processing system described in Appendix 1.

[0118] (Note 3) The aforementioned probability parameter is the inverse of the variance-covariance matrix, and is partially constrained to zero. The signal processing system described in Appendix 1 or Appendix 2.

[0119] (Note 4) The aforementioned complex vector has as many elements as there are inputs obtained from each corresponding pixel when multiple aligned complex images are taken as input. A signal processing system as described in any of the appendices 1 to 3.

[0120] (Note 5) The parameter calculation means calculates the probability parameter using a prior distribution for the probability parameter. A signal processing system as described in any of the appendices 1 through 4.

[0121] (Note 6) The system includes an output means that outputs at least one of the following: information indicating pixels assigned to a cluster and information indicating the probability parameters of the cluster. A signal processing system as described in any of the appendices 1 through 5.

[0122] (Note 7) The system includes interferometric analysis means for performing analysis of the cluster using the aforementioned probability parameters of the cluster. A signal processing system as described in any of the appendices 1 through 6.

[0123] (Note 8) The system includes a display means that displays the results of cluster analysis by the interferometry analysis means at the pixel positions assigned to the cluster. A signal processing system as described in any of the appendices 1 through 7.

[0124] (Note 9) Computers Each pixel is assigned to a cluster such that the complex measure indicating the degree of match between the complex vectors obtained from corresponding pixels of multiple complex images and the cluster's probability parameters is maximized. For each cluster, the probability parameter is calculated such that the complex scale is highest for the entire set of pixels to which it is assigned. The aforementioned allocation and calculation are performed alternately. A signal processing method characterized by the following:

[0125] (Note 10) On the computer, An assignment process that assigns each pixel to a cluster such that a complex measure indicating the degree of match between a complex number vector obtained from corresponding pixels of multiple complex images and the cluster's probability parameter is maximized, For each cluster, a parameter calculation process is performed to calculate the probability parameter such that the complex scale is highest for the entire set of pixels assigned to it. The assignment process and the parameter calculation process are executed alternately. A signal processing program to be executed.

[0126] (Note 11) A non-temporary computer-readable recording medium on which a signal processing program is stored, The aforementioned signal processing program is provided to the computer, An assignment process that assigns each pixel to a cluster such that a complex measure indicating the degree of match between a complex number vector obtained from corresponding pixels of multiple complex images and the cluster's probability parameter is maximized, For each cluster, a parameter calculation process is performed to calculate the probability parameter such that the complex scale is highest for the entire set of pixels assigned to it. The assignment process and the parameter calculation process are executed alternately. A non-temporary, computer-readable recording medium.

[0127] Some or all of the elements (e.g., configuration and function) described in Appendices 2 to 8 that are dependent on Appendice 1 may also be dependent on Appendices 9, 10, and 11 in the same way as those described in Appendices 2 to 8. Some or all of the elements described in any appendice may be applicable to various hardware, software, recording means, systems, and methods for recording software. [Explanation of Symbols]

[0128] 10,400,401 Signal Processing System 11. Allocation means 12 Parameter calculation means 100,101 Signal Processing Device 110 pixel allocation section 120,121 Parameter calculation section 130 Assignment Information Storage Unit 140 Parameter Information Storage Unit 150 Output section 200 SAR image storage unit 210 Pre-distribution storage unit 410 Displacement Analysis Unit 420,421 Display section 430 Change detection unit 1000 computers 1001 Processor 1002 Storage device 1003 memory 1004 Bus 1005 Input / Output Interface 1006 Network Interface

Claims

1. An assignment means for assigning each pixel to a cluster such that a complex measure indicating the degree of match between a complex number vector obtained from corresponding pixels of multiple complex images and the probability parameter of the cluster is maximized, For each cluster, the system includes a parameter calculation means for calculating the probability parameter such that the complex scale is highest for the entire set of pixels assigned to it. The assignment means and the parameter calculation means operate alternately. A signal processing system characterized by the following:

2. The aforementioned complex scale is a zero-mean multivariate complex Gaussian distribution, The aforementioned probability parameter is the variance-covariance matrix of the multivariate complex Gaussian distribution or the inverse matrix of said variance-covariance matrix. The signal processing system according to claim 1.

3. The aforementioned probability parameter is the inverse of the variance-covariance matrix, and is partially constrained to zero. The signal processing system according to claim 1 or claim 2.

4. The aforementioned complex vector has as many elements as there are inputs obtained from each corresponding pixel when multiple aligned complex images are taken as input. Signal processing system according to claim 1 or claim 2

5. The parameter calculation means calculates the probability parameter using a prior distribution for the probability parameter. The signal processing system according to claim 1 or claim 2.

6. The system includes an output means that outputs at least one of the following: information indicating pixels assigned to a cluster and information indicating the probability parameters of the cluster. The signal processing system according to claim 1 or claim 2.

7. The system includes interferometric analysis means for performing analysis of the cluster using the aforementioned probability parameters of the cluster. The signal processing system according to claim 1 or claim 2.

8. The system includes a display control means that displays the cluster analysis results obtained by the interferometry analysis means at the pixel positions assigned to the cluster. The signal processing system according to claim 7.

9. Computers Each pixel is assigned to a cluster such that the complex measure indicating the degree of match between the complex vectors obtained from corresponding pixels of multiple complex images and the cluster's probability parameters is maximized. For each cluster, the probability parameter is calculated such that the complex scale is highest for the entire set of pixels to which it is assigned. The aforementioned allocation and calculation are performed alternately. A signal processing method characterized by the following:

10. On the computer, An assignment process that assigns each pixel to a cluster such that a complex measure indicating the degree of match between a complex number vector obtained from corresponding pixels of multiple complex images and the cluster's probability parameter is maximized, For each cluster, a parameter calculation process is performed to calculate the probability parameter such that the complex scale is highest for the entire set of pixels assigned to it. The assignment process and the parameter calculation process are executed alternately. A signal processing program for this purpose.

Images