A method and system for detecting human activity changes based on SAR images

Abstract

The present invention discloses a method and system for detecting changes in human activities based on SAR images, and relates to the field of ecology and remote sensing information technology. The method includes: obtaining SAR images of the target area at different periods, and spatially registering the SAR images of the different periods based on DEM data; performing change detection on the spatially registered SAR images based on multi-scale deep feature fusion to obtain a SAR image change detection map; comparing and analyzing the SAR image change detection map with a human activity feature data set to obtain suspected ecological damage spots; modifying the boundaries and assigning attributes to the suspected ecological damage spots to obtain the final ecological damage spots. The present invention solves the problem of insufficient coverage of optical images in cloudy and rainy areas, expands the application of SAR images in the field of human activity supervision, and can meet the all-round, high-precision, and short-cycle supervision needs of ecological protection red lines.

Description

Technical Field

[0001] The present invention relates to the field of ecology and remote sensing information technology, and more specifically to a method and system for detecting changes in human activities based on SAR images. Background Art

[0002] At present, the supervision of human activities within the ecological protection red line mainly adopts manual visual interpretation and automatic change detection based on high-resolution optical remote sensing images. However, optical images are greatly affected by weather. In cloudy and rainy areas, there are often problems such as difficulty in obtaining optical data or poor image quality, making it difficult to ensure the accuracy and timeliness of remote sensing monitoring data results.

[0003] Therefore, how to solve the problems of low accuracy and poor timeliness in monitoring changes in human activities in optical remote sensing images in cloudy and rainy areas, and to construct a radar and optical spatiotemporal coupling method for identifying typical human interference targets in cloudy and rainy areas are issues that technical personnel in this field urgently need to solve. Summary of the invention

[0004] In view of this, the present invention provides a method and system for detecting human activity changes based on SAR images, which are used to solve the problems of low accuracy and poor timeliness in monitoring human activity changes using optical remote sensing images in cloudy and rainy areas.

[0005] In order to achieve the above object, the present invention adopts the following technical solution:

[0006] The present invention first discloses a method for detecting changes in human activities based on SAR images. Figure 1 As shown, the following steps are included:

[0007] Acquire SAR images of the target area at different times, and perform spatial registration on the SAR images at different times based on DEM data;

[0008] Based on multi-scale deep feature fusion, change detection is performed on the spatially registered SAR image to obtain a SAR image change detection map;

[0009] Compare and analyze the SAR image change detection map with the human activity feature dataset to obtain suspected ecological damage patches;

[0010] The boundaries of suspected ecological damage spots are modified and attributes are assigned to obtain the final ecological damage spots.

[0011] Furthermore, spatial registration of the SAR images of the different periods is performed based on the DEM data, specifically comprising the following steps:

[0012] S1.1: Acquire two SAR images S1 and S2 of the target area at different times;

[0013] S1.2: Obtain DEM data within the target area and select control points of the DEM data;

[0014] S1.3: transferring the control points of the DEM data to the simulated image;

[0015] S1.4: Establish a polynomial transformation between the simulated image and each SAR image;

[0016] S1.5: Use the point set matching algorithm based on vector field consistency to transfer the control points of the simulated image to each SAR image;

[0017] S1.6: Establish the polynomial transformation model of the two SAR images S1 and S2;

[0018] S1.7. Resampling is performed to realize spatial registration of SAR images, and two spatially registered SAR images I1 and I2 are obtained.

[0019] Furthermore, in step S1.3, the control points of the DEM data are transferred to the simulated image, specifically including:

[0020] The range-Doppler RD model is used to establish a lookup table for associating DEM data with simulated images, namely: In the formula, (X, Y, Z) represents the spatial coordinates of DEM; (i, j) represents the spatial coordinates of the simulated image;

[0021] Match the simulated image with the SAR image to obtain the coordinate correspondence between the simulated image SIM and the SAR image, namely: Where (i1, j1) represents the spatial coordinates of the SAR image.

[0022] Furthermore, a polynomial transformation between the simulated image and each SAR image is established, including the following polynomial transformations:

[0023] i1=a0+a1i+a2j+a3ij+a4i 2 +a5j 2 ;

[0024] j1=b0+b1i+b2j+b3ij+b4i 2 +b5j 2 ;

[0025] In the formula, a0, a1, a2, a3, a4, a5, b0, b1, b2, b3, b4, and b5 all represent polynomial coefficients.

[0026] Furthermore, a polynomial transformation model of the two SAR images S1 and S2 is established, including the following polynomials:

[0027] X S1 =a0+a1X S2 +a2Y S2 +a3X S2 Y S2 +a4X S2 2 +a5Y S2 2 ;

[0028] Y S2 =b0+b1X S1 +b2Y S1 +b3X S1 Y S1 +b4X S1 2 +b5Y S1 2 ;

[0029] Where, X S2 ,Y S2 Indicates the coordinates of the feature points of the S2 image; X S1 ,Y S1 represents the coordinates of the feature points of the S1 image; a0, a1, a2, a3, a4, a5, b0, b1, b2, b3, b4, b5 all represent the polynomial coefficients in step S1.4.

[0030] Furthermore, change detection is performed on the spatially registered SAR image based on multi-scale deep feature fusion, which specifically includes the following steps:

[0031] S2.1: Perform logarithmic operation on the spatially registered SAR images I1 and I2 to obtain a logarithmic ratio difference map;

[0032] S2.2: performing three-layer synchronous squeezing wavelet transform decomposition on the logarithmic ratio difference map obtained in step S2.1 to obtain low-frequency components and high-frequency components after each layer of decomposition;

[0033]

[0034]

[0035]

[0036] Where: I LR It represents the logarithmic ratio difference map obtained by performing logarithmic operation on the SAR images I1 and I2 after spatial registration; Represents the low-frequency components of the first layer; Represents the three high-frequency components of the first layer; represents the low-frequency component of the second layer; represents the three high-frequency components of the second layer; represents the low-frequency components of the third layer; Represents the three high-frequency components of the third layer;

[0037] S2.3: Based on the low-frequency components and high-frequency components after decomposition of each layer, an inverse two-dimensional static wavelet transform is performed to obtain a multi-scale difference map through independent reconstruction.

[0038] S2.4: The multi-scale difference map is divided into two categories using the fuzzy cluster analysis FCM algorithm, the classified multi-scale difference map is used as a pseudo label for training CNN, and reliable training samples are selected from the multi-scale difference map;

[0039] S2.5: Input the training samples into the CNN model, use the BP algorithm of stochastic gradient descent to train the CNN, and use the trained CNN model to generate change detection result images of two SAR images I1 and I2.

[0040] Furthermore, in step S2.1, logarithmic operation is performed on the spatially registered SAR images I1 and I2, specifically including the following expression:

[0041] I LR (x,y)=∣logI2(x,y)-logI1(x,y)∣;

[0042] In the formula, I LR (x, y) represents the logarithmic ratio difference image element value, I1(x, y) and I2(x, y) are the pixel values ​​corresponding to the I1 and I2 images respectively.

[0043] Furthermore, in step S2.3, the inverse two-dimensional static wavelet transform is specifically implemented by the following expression:

[0044]

[0045]

[0046] Where: represents the reconstructed low-frequency component of the third layer; represents the reconstructed high-frequency components of the third layer; represents the reconstructed low-frequency component of the second layer;

[0047] represents the reconstructed high-frequency components of the second layer; Represents the reconstructed low-frequency components of the first layer.

[0048] Furthermore, the human activity feature dataset is obtained by the following steps:

[0049] Step 1: Obtain typical ground feature samples, which include six types of development and construction: mineral resource development, industrial development and construction, energy development and construction, tourism development and construction, transportation development and construction, and other development and construction types other than the above typical ground feature samples;

[0050] Step 2: Analyze the optical image features of the typical ground object samples, wherein the optical image features include spatial features, attribute features and texture features;

[0051] Step 3: Analyze the SAR image features of the typical ground object samples, wherein the SAR image features include imaging scattering, geometric features and basic image features;

[0052] Step 4: Analyze the performance characteristics of the typical ground object sample target on the SAR image from the perspective of color tone, texture, geometric shape characteristics and context characteristics;

[0053] Step 5: Match the optical image of the typical ground object sample with the corresponding SAR image;

[0054] Step 6: Calculate eight characteristic statistical parameters of the target object sample based on the gray level co-occurrence matrix, wherein the eight characteristic statistical parameters include entropy, mean, variance, contrast, correlation, dissimilarity, homogeneity and angular second moment;

[0055] Step 7: Use eight characteristic statistical parameters and the grayscale characteristics of the target object's optical image to jointly construct a characteristic data set of the target object.

[0056] In addition, the present invention also discloses a human activity change detection system based on SAR images, comprising a plurality of interconnected computer modules, which can implement any one of the human activity change detection methods based on SAR images described in the present invention when running.

[0057] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses a method and system for detecting changes in human activities based on SAR images, which has the following beneficial effects:

[0058] The present invention solves the problem of insufficient coverage of optical images in cloudy and rainy areas by establishing an intelligent automatic extraction system for changes in human activities along the ecological protection red line that coordinates optical and SAR images, expands the application of SAR images in the field of human activity supervision, and can meet the needs of all-round, high-precision, and short-cycle supervision of the ecological protection red line. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0060] Figure 1 A schematic diagram of the overall process of a method for detecting changes in human activities based on SAR images provided in an embodiment of the present invention.

[0061] Figure 2 A flow chart of registering two SAR images based on simulated images provided in an embodiment of the present invention.

[0062] Figure 3 The logarithmic ratio difference provided by the embodiment of the present invention Figure 3 Schematic diagram of layer decomposition and reconstruction.

[0063] Figure 4 This is a flow chart of a SAR image change detection algorithm provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0064] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] The embodiment of the present invention discloses a method for detecting changes in human activities based on SAR images, such as Figure 1 As shown, the following steps are included:

[0066] S1. Obtain SAR images of the target area at different times, and perform spatial registration of SAR images at different times based on DEM data;

[0067] S2, performing change detection on the spatially registered SAR image based on multi-scale deep feature fusion to obtain a SAR image change detection map;

[0068] S3, comparing and analyzing the SAR image change detection map with the human activity feature data set to obtain suspected ecological damage spots;

[0069] S4. Modify the boundaries and assign attributes to the suspected ecological damage spots to obtain the final ecological damage spots.

[0070] Each step is further described below.

[0071] like Figure 2 As shown, in step S1, spatial registration of the SAR image is performed based on DEM data, which may specifically include the following steps:

[0072] S1.1: Acquire two SAR images S1 and S2 of the target area at different times;

[0073] S1.2: Obtain DEM data within the target area and select control points of the DEM data;

[0074] S1.3: Transferring the control points of the DEM data to the simulated image, specifically including:

[0075] The range-Doppler RD model is used to establish a lookup table for associating DEM data with simulated images, namely: In the formula, (X, Y, Z) represents the spatial coordinates of DEM; (i, j) represents the spatial coordinates of the simulated image;

[0076] Match the simulated image with the SAR image to obtain the coordinate correspondence between the simulated image SIM and the SAR image, namely: Where, (i1, j1) represents the spatial coordinates of the SAR image;

[0077] S1.4: Establish a polynomial transformation between the simulated image and each SAR image;

[0078] i1=a0+a1i+a2j+a3ij+a4i 2 +a5j 2 ;

[0079] j1=b0+b1i+b2j+b3ij+b4i 2 +b5j 2 ;

[0080] In the formula, a0, a1, a2, a3, a4, a5, b0, b1, b2, b3, b4, b5 all represent polynomial coefficients;

[0081] S1.5: Use the vector field consistency point set matching (VFC) algorithm to transfer the control points of the simulated image to each SAR image;

[0082] S1.6: Establish the polynomial transformation model of the two SAR images S1 and S2;

[0083] X S1 =a0+a1X S2 +a2Y S2 +a3X S2 Y S2 +a4X S2 2+a5Y S2 2 ;

[0084] Y S2 =b0+b1X S1 +b2Y S1 +b3X S1 Y S1 +b4X S1 2 +b5Y S1 2 ;

[0085] Where, X S2 ,Y S2 Indicates the coordinates of the feature points of the S2 image; X S1 ,Y S1 represents the coordinates of the feature points of the S1 image; a0, a1, a2, a3, a4, a5, b0, b1, b2, b3, b4, b5 all represent the polynomial coefficients in step S1.4.

[0086] S1.7. Resampling is performed to realize spatial registration of SAR images, and two spatially registered SAR images I1 and I2 are obtained.

[0087] In step S2, change detection is performed on the spatially registered SAR image based on multi-scale deep feature fusion, which may specifically include the following steps:

[0088] S2.1: Perform logarithmic operation on the spatially registered SAR images I1 and I2 to obtain a logarithmic ratio difference map.

[0089] The logarithmic ratio difference can be obtained by the following formula:

[0090] I LR (x,y)=∣logI2(x,y)-logI1(x,y)∣; where, I LR (x,y) means

[0091] Logarithmic ratio difference image element value, I1(x,y) and I2(x,y) are the pixel values ​​corresponding to the I1 and I2 images respectively.

[0092] S2.2: performing three-layer synchronous squeezing wavelet transform decomposition on the logarithmic ratio difference map obtained in step S2.1 to obtain low-frequency components and high-frequency components after each layer of decomposition;

[0093]

[0094]

[0095]

[0096] Where: ILR It represents the logarithmic ratio difference map obtained by performing logarithmic operation on the SAR images I1 and I2 after spatial registration; Represents the low-frequency components of the first layer; Represents the three high-frequency components of the first layer; represents the low-frequency component of the second layer; represents the three high-frequency components of the second layer; represents the low-frequency components of the third layer; Represents the three high-frequency components of the third layer.

[0097] S2.3: Based on the low-frequency components and high-frequency components after each layer of decomposition, an inverse two-dimensional static wavelet transform is performed to obtain a multi-scale difference map through independent reconstruction; the inverse two-dimensional static wavelet transform is specifically implemented by the following expression:

[0098]

[0099]

[0100] Where: represents the reconstructed low-frequency component of the third layer; represents the reconstructed high-frequency components of the third layer; represents the reconstructed low-frequency component of the second layer;

[0101] represents the reconstructed high-frequency components of the second layer; Represents the reconstructed low-frequency components of the first layer.

[0102] The log ratio difference in the above process Figure 3 The overall process of layer decomposition and reconstruction is as follows Figure 3 As shown, Figure 4 A flow chart of a SAR image change detection algorithm provided in an embodiment.

[0103] In this embodiment, the human activity feature dataset is obtained by the following steps:

[0104] Step 1: Obtain typical ground feature samples, which include six types of development and construction: mineral resource development, industrial development and construction, energy development and construction, tourism development and construction, transportation development and construction, and other development and construction types other than the above typical ground feature samples;

[0105] Step 2: Analyze the optical image features of the typical ground object samples, wherein the optical image features include spatial features, attribute features and texture features;

[0106] Step 3: Analyze the SAR image features of the typical ground object samples, wherein the SAR image features include imaging scattering, geometric features and basic image features;

[0107] Step 4: Analyze the performance characteristics of the typical ground object sample target on the SAR image from the perspective of color tone, texture, geometric shape characteristics and context characteristics;

[0108] Step 5: Match the optical image of the typical ground object sample with the corresponding SAR image;

[0109] Step 6: Calculate eight characteristic statistical parameters of the target object sample based on the gray level co-occurrence matrix, wherein the eight characteristic statistical parameters include entropy, mean, variance, contrast, correlation, dissimilarity, homogeneity and angular second moment;

[0110] Step 7: Use eight characteristic statistical parameters and the grayscale characteristics of the target object's optical image to jointly construct a characteristic data set of the target object.

[0111] In addition, an embodiment of the present invention further discloses a human activity change detection system based on SAR images, comprising a plurality of interconnected computer modules, wherein when the computer modules are running, any one of the human activity change detection methods based on SAR images described in the present invention can be implemented.

[0112] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same and similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts refer to the method part description. The above description of the disclosed embodiments enables professionals and technicians in this field to implement or use the present invention. Various modifications to these embodiments will be obvious to professionals and technicians in this field, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown in this article, but will comply with the widest range consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting human activity changes based on SAR images, characterized in that: The following steps are involved: Acquire SAR images of the target area at different times, and perform spatial registration on the SAR images at different times based on DEM data, specifically including the following steps: S1.1: Acquire two SAR images S1 and S2 of the target area at different times; S1.2: Obtain DEM data within the target area and select control points of the DEM data; S1.3: Transferring the control points of the DEM data to the simulated image, specifically including: The range-Doppler RD model is used to establish a lookup table for associating DEM data with simulated images, namely: Where (X, Y, Z) represents the spatial coordinates of the DEM; (i, j) represents the spatial coordinates of the simulated image; Match the simulated image with the SAR image to obtain the coordinate correspondence between the simulated image SIM and the SAR image, namely: Where, (i1, j1) represents the spatial coordinates of the SAR image; S1.4: Establish a polynomial transformation between the simulated image and each SAR image; The following polynomial transformations are included: i1=a0+a1i+a2j+a3ij+a4i 2 +a5j 2 ; j1=b0+b1i+b2j+b3ij+b4i 2 +b5j 2 ; In the formula, a0, a1, a2, a3, a4, a5, b0, b1, b2, b3, b4, b5 all represent polynomial coefficients; S1.5: Use the point set matching algorithm based on vector field consistency to transfer the control points of the simulated image to each SAR image; S1.6: Establish a polynomial transformation model for the two SAR images S1 and S2, including the following polynomials: <h2 style=";text-align:left;direction:ltr">X<h2 style=";text-align:left;direction:ltr"> S1 <h2 style=";text-align:left;direction:ltr"> =a0+a1X<h2 style=";text-align:left;direction:ltr"> S2 <h2 style=";text-align:left;direction:ltr"> +a2Y<h2 style=";text-align:left;direction:ltr"> S2 <h2 style=";text-align:left;direction:ltr"> +a3X<h2 style=";text-align:left;direction:ltr"> S2 <h2 style=";text-align:left;direction:ltr"> Y<h2 style=";text-align:left;direction:ltr"> S2 <h2 style=";text-align:left;direction:ltr"> +a4X<h2 style=";text-align:left;direction:ltr"> S2 <h2 style=";text-align:left;direction:ltr"> 2 <h2 style=";text-align:left;direction:ltr"> +a5Y<h2 style=";text-align:left;direction:ltr"> S2 <h2 style=";text-align:left;direction:ltr"> 2 <h2 style=";text-align:left;direction:ltr"> ; <h2 style=";text-align:left;direction:ltr">Y<h2 style=";text-align:left;direction:ltr"> S2 <h2 style=";text-align:left;direction:ltr"> =b0+b1X<h2 style=";text-align:left;direction:ltr"> S1 <h2 style=";text-align:left;direction:ltr"> +b2Y<h2 style=";text-align:left;direction:ltr"> S1 <h2 style=";text-align:left;direction:ltr"> +b3X<h2 style=";text-align:left;direction:ltr"> S1 <h2 style=";text-align:left;direction:ltr"> Y<h2 style=";text-align:left;direction:ltr"> S1 <h2 style=";text-align:left;direction:ltr"> +b4X<h2 style=";text-align:left;direction:ltr"> S1 <h2 style=";text-align:left;direction:ltr"> 2 <h2 style=";text-align:left;direction:ltr"> +b5Y<h2 style=";text-align:left;direction:ltr"> S1 <h2 style=";text-align:left;direction:ltr"> 2 <h2 style=";text-align:left;direction:ltr"> ; Where, X S2 ,Y S2 Indicates the coordinates of the feature points of the S2 image; X S1 , Y S1 Represents the coordinates of the feature points of the S1 image; a0, a1, a2, a3, a4, a5, b0, b1, b2, b3, b4, b5 all represent the polynomial coefficients in step S1.4; S1.7, resampling to achieve spatial registration of SAR images, and obtain two spatially registered SAR images I1 and I2; Based on multi-scale deep feature fusion, change detection is performed on the spatially registered SAR image to obtain a SAR image change detection map; Compare and analyze the SAR image change detection map with the human activity feature data set to obtain suspected ecological damage spots; the human activity feature data set is obtained by the following steps: Step 1: Obtain typical ground feature samples, which include six types of development and construction: mineral resource development, industrial development and construction, energy development and construction, tourism development and construction, transportation development and construction, and other development and construction except the above typical ground feature samples; Step 2: Analyze the optical image features of the typical ground object samples, wherein the optical image features include spatial features, attribute features and texture features; Step 3: Analyze the SAR image features of the typical ground object samples, wherein the SAR image features include imaging scattering, geometric features and basic image features; Step 4: Analyze the performance characteristics of the typical ground sample target on the SAR image from the perspective of color tone, texture, geometric shape characteristics and context characteristics; Step 5: Match the optical image of the typical ground object sample with the corresponding SAR image; Step 6: Calculate eight characteristic statistical parameters of the target object sample based on the gray level co-occurrence matrix, wherein the eight characteristic statistical parameters include entropy, mean, variance, contrast, correlation, dissimilarity, homogeneity and angular second moment; Step 7: Use eight characteristic statistical parameters and the grayscale characteristics of the target object's optical image to construct a characteristic data set of the target object; The suspected ecological damage spots are modified in terms of boundaries and attributes to obtain final ecological damage spots.

2. The method for detecting changes in human activities based on SAR images according to claim 1, characterized in that: Change detection is performed on the spatially registered SAR image based on multi-scale deep feature fusion, which specifically includes the following steps: S2.1: Perform logarithmic operation on the spatially registered SAR images I1 and I2 to obtain a logarithmic ratio difference map; S2.2: performing three-layer synchronous squeezing wavelet transform decomposition on the logarithmic ratio difference map obtained in step S2.1 to obtain low-frequency components and high-frequency components after each layer of decomposition; Where: I LR It represents the logarithmic ratio difference map obtained by performing logarithmic operation on the SAR images I1 and I2 after spatial registration; Represents the low-frequency components of the first layer; Represents the three high-frequency components of the first layer; represents the low-frequency component of the second layer; represents the three high-frequency components of the second layer; represents the low-frequency components of the third layer; Represents the three high-frequency components of the third layer; S2.3: Based on the low-frequency components and high-frequency components after each layer of decomposition, perform inverse two-dimensional static wavelet transform. Obtain multi-scale difference maps through independent reconstruction; S2.4: The multi-scale difference map is divided into two categories using the fuzzy cluster analysis FCM algorithm, the classified multi-scale difference map is used as a pseudo label for training CNN, and reliable training samples are selected from the multi-scale difference map; S2.5: Input the training samples into the CNN model, train the CNN using the BP algorithm of stochastic gradient descent, and use the trained CNN model to generate change detection result images of two SAR images I1 and I2.

3. The method for detecting changes in human activities based on SAR images according to claim 2, characterized in that: In step S2.1, logarithmic operation is performed on the spatially registered SAR images I1 and I2, specifically including the following expression: I LR (x,y)=|logI2(x,y)-logI1(x,y)|; In the formula, I LR (x, y) represents the logarithmic ratio difference image element value, I1(x, y) and I2(x, y) are spatial The pixel values ​​corresponding to the registered SAR images I1 and I2.

4. The method for detecting changes in human activities based on SAR images according to claim 2, characterized in that: In step S2.3, the inverse two-dimensional static wavelet transform is specifically implemented by the following expression: Where: represents the reconstructed low-frequency component of the third layer; represents the reconstructed high-frequency components of the third layer; represents the reconstructed low-frequency component of the second layer; represents the reconstructed high-frequency components of the second layer; Represents the reconstructed low-frequency components of the first layer.

5. A human activity change detection system based on SAR images, characterized in that: It comprises a plurality of interconnected computer modules, and when the computer modules are running, the method for detecting changes in human activities based on SAR images as described in any one of claims 1 to 4 can be implemented.

Images