A method and system for monitoring human settlement environment based on remote sensing

Abstract

The present invention provides a method and system for monitoring a human settlement environment based on remote sensing, which relates to the field of data processing technology. In the present invention, images of a target human settlement environment are collected and multiple remote sensing image sets are output. An image processing neural network is used to screen the multiple human settlement environment remote sensing images included in each remote sensing image set, so as to output a target remote sensing image set corresponding to each remote sensing image set, and each target remote sensing image set includes multiple human settlement environment remote sensing images. Based on the included human settlement environment remote sensing images, a correlation analysis is performed on the output multiple target remote sensing image sets to output a target environmental change analysis result of the target human settlement environment, which is used to reflect the degree of environmental change of the target human settlement environment. Based on the above method, the reliability of human settlement environment monitoring can be improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method and system for monitoring a human settlement environment based on remote sensing. Background Art

[0002] The development of remote sensing technology has led to a gradual increase in its application scenarios. For example, in human settlement environment monitoring, remote sensing images can be analyzed and processed to obtain corresponding monitoring results. However, different monitoring requirements require different analysis and processing methods for remote sensing images. For example, some requirements may require correlation analysis of remote sensing images to output environmental change analysis results, namely the degree of environmental change. However, existing technologies have the problem of low reliability of correlation analysis, which means that the reliability of environmental monitoring is low. Summary of the Invention

[0003] In view of this, an object of the present invention is to provide a method and system for monitoring a human living environment based on remote sensing, so as to improve the reliability of human living environment monitoring.

[0004] To achieve the above objectives, the embodiments of the present invention adopt the following technical solutions:

[0005] A method for monitoring a human settlement environment based on remote sensing, applied to an environment monitoring server, includes:

[0006] Capturing images of a target human settlement environment and outputting a plurality of remote sensing image sets, each of the remote sensing image sets comprising a plurality of remote sensing images of the human settlement environment, wherein the time interval between image acquisition times corresponding to each two remote sensing image sets satisfies a preset time interval condition;

[0007] Using an image processing neural network, screening the multiple human settlement environment remote sensing images included in each of the remote sensing image sets to output a target remote sensing image set corresponding to each of the remote sensing image sets, each of the target remote sensing image sets including the multiple human settlement environment remote sensing images;

[0008] Based on the included human settlement environment remote sensing images, correlation analysis is performed on the output multiple target remote sensing image sets to output the target environment change analysis results of the target human settlement environment, and the target environment change analysis results are used to reflect the degree of environmental change of the target human settlement environment.

[0009] In some preferred embodiments, in the above-mentioned remote sensing-based human settlement environment monitoring method, the step of collecting images of the target human settlement environment and outputting a plurality of remote sensing image sets includes:

[0010] After performing any image acquisition of the target human settlement environment, outputting a remote sensing image set corresponding to the image acquisition and determining acquisition parameters of the image acquisition;

[0011] Before performing a next image acquisition on the target human living environment, statistical processing of environmental events is performed on the target human living environment to output a statistical value of environmental events currently corresponding to the target human living environment, wherein the environmental events are used to reflect events that occur in time periods corresponding to two corresponding image acquisitions and cause changes to the target human living environment, and the next image acquisition refers to an image acquisition performed after any of the image acquisitions;

[0012] Adjusting acquisition parameters corresponding to any one image acquisition according to the environmental event statistics to output target acquisition parameters corresponding to the next image acquisition;

[0013] The next image acquisition is performed on the target human settlement environment according to the target acquisition parameters to output a remote sensing image set corresponding to the image acquisition.

[0014] In some preferred embodiments, in the above-mentioned remote sensing-based human settlement environment monitoring method, the step of using an image processing neural network to screen the multiple human settlement environment remote sensing images included in each remote sensing image set to output a target remote sensing image set corresponding to each remote sensing image set includes:

[0015] Performing image segmentation on the remote sensing image set to form an image segmentation segment set, the image segmentation segment set including a first remote sensing image segment and a number greater than or equal to one second remote sensing image segment, the first remote sensing image segment including at least one human settlement remote sensing image, each of the second remote sensing image segments including at least one human settlement remote sensing image, when the first remote sensing image segment includes a plurality of human settlement remote sensing images, the plurality of human settlement remote sensing images are continuous in image time sequence, and when the second remote sensing image segment includes a plurality of human settlement remote sensing images, the plurality of human settlement remote sensing images are continuous in image time sequence;

[0016] Loading the first remote sensing image segments and the second remote sensing image segments greater than or equal to one into an image processing neural network, processing the first remote sensing image segments and the second remote sensing image segments greater than or equal to one using the image processing neural network, and outputting image segment close relationships and close relationship label information corresponding to each of the second remote sensing image segments greater than or equal to one, wherein each image segment close relationship is used to reflect whether the corresponding second remote sensing image segment and the first remote sensing image segment are matching image segments;

[0017] splicing an image segment relationship network based on the image segment close relationship and the close relationship label information corresponding to each second remote sensing image segment, wherein the image segment close relationship reflects whether there is a network connection line between corresponding positions in the image segment relationship network, and the close relationship label information reflects the connection line characteristics of the network connection line;

[0018] Wandering in the image segment relationship network to form a network link set, and performing link screening on the network link set to output target network relationship links corresponding to the image segmentation segment set, wherein each network relationship link included in the network link set includes the first remote sensing image segment, and any second remote sensing image segment in each network relationship link has a network connection line with the first remote sensing image segment in the image segment relationship network;

[0019] According to the first remote sensing image segment and each second remote sensing image segment included in each target network relationship link, a target remote sensing image set corresponding to the remote sensing image set is screened out.

[0020] In some preferred embodiments, in the above-mentioned remote sensing-based human settlement environment monitoring method, the step of loading the first remote sensing image segment and the second remote sensing image segment whose number is greater than or equal to 1 into an image processing neural network, processing the first remote sensing image segment and the second remote sensing image segment whose number is greater than or equal to 1 by using the image processing neural network, and outputting the image segment close relationship and close relationship label information corresponding to each second remote sensing image segment whose number is greater than or equal to 1 includes:

[0021] Loading the first remote sensing image segment and the second remote sensing image segment whose number is greater than or equal to 1 into an image processing neural network, processing the first remote sensing image segment and the second remote sensing image segment whose number is greater than or equal to 1 using the image processing neural network, outputting an image variable mapping continuous vector corresponding to the first remote sensing image segment, and outputting an image variable mapping continuous vector and an image meaning vector corresponding to each of the second remote sensing image segments whose number is greater than or equal to 1;

[0022] performing random pairing processing on the first remote sensing image segment and the second remote sensing image segment greater than or equal to 1 according to an image time sequence relationship of the remote sensing image segments included in the image segmentation segment set, and outputting corresponding image segment pairing combinations greater than or equal to 2;

[0023] Using the image processing neural network, perform image vector splicing processing on the image segment pairing combinations greater than or equal to 2, and output image segment close relationships and close relationship label information corresponding to each image segment pairing combination;

[0024] Based on the image segment close relationships and close relationship label information corresponding to the image segment pairing combinations greater than or equal to 2, the image segment close relationships and close relationship label information corresponding to the second remote sensing image segments greater than or equal to 1 are determined.

[0025] In some preferred embodiments, in the above-mentioned remote sensing-based human settlement environment monitoring method, the step of using the image processing neural network to perform image vector splicing processing on the image segment pairing combinations greater than or equal to 2, and outputting the image segment close relationship and close relationship label information corresponding to each image segment pairing combination, includes:

[0026] Using an image vector splicing unit included in the image processing neural network, the image variable mapping continuous vector and the image meaning vector corresponding to the second remote sensing image segment included in each of the image segment pairing combinations are subjected to image vector splicing processing with the image variable mapping continuous vector corresponding to the first remote sensing image segment, and the image splicing vector corresponding to each of the image segment pairing combinations is output;

[0027] Using the discriminant output function included in the image processing neural network, vector discrimination is performed on each of the image splicing vectors, and the image segment close relationships and close relationship label information corresponding to the image segment pairing combinations greater than or equal to 2 are output.

[0028] In some preferred embodiments, in the above-mentioned remote sensing-based human settlement environment monitoring method, the step of determining the image segment close relationships and close relationship label information corresponding to the second remote sensing image segments greater than or equal to 1 based on the image segment close relationships and close relationship label information corresponding to the image segment pairing combinations greater than or equal to 2, includes:

[0029] Each of the image segment pairing combinations whose number is greater than or equal to 2 is processed by the following steps in sequence or simultaneously:

[0030] If the image segment pairing combination consists of a second remote sensing image segment, marking the image segment close relationship and close relationship label information corresponding to the image segment pairing combination as the image segment close relationship and close relationship label information corresponding to the second remote sensing image segment;

[0031] If the image segment pairing combination is composed of a plurality of second remote sensing image segments, the image segment close relationship and close relationship label information corresponding to the image segment pairing combination are marked as the image segment close relationship and close relationship label information corresponding to the second remote sensing image segment that is the earliest in image time sequence among the plurality of second remote sensing image segments;

[0032] Based on the image segment close relationship and close relationship label information corresponding to the one second remote sensing image segment, or based on the image segment close relationship and close relationship label information corresponding to the second remote sensing image segment at the front of the image time sequence, determine the image segment close relationship and close relationship label information corresponding to the second remote sensing image segments whose number is greater than or equal to 1.

[0033] In some preferred embodiments, in the above-mentioned remote sensing-based human settlement environment monitoring method, the step of splicing to form an image segment relationship network based on the image segment close relationship and the close relationship label information corresponding to each second remote sensing image segment includes:

[0034] Marking the network distribution information of the second remote sensing image segments and the first remote sensing image segments respectively having a number greater than or equal to 1, based on the image segment close relationship and the close relationship label information corresponding to each second remote sensing image segment;

[0035] Based on the network distribution information corresponding to each of the second remote sensing image segments and the network distribution information corresponding to the first remote sensing image segment, a connection direction with the second remote sensing image segment as the connection starting point and the first remote sensing image segment as the connection end point is used to splice the second remote sensing image segments and the first remote sensing image segments whose number is greater than or equal to 1 to form an image segment relationship network.

[0036] In some preferred embodiments, in the above-mentioned remote sensing-based human settlement environment monitoring method, the step of performing link screening on the network link set and outputting the target network relationship links corresponding to the image segmentation set includes:

[0037] Perform screening processing on each of the network relationship links included in the network link set by sequentially or synchronously performing the following steps:

[0038] Identifying the network relationship link to determine whether the network relationship link includes the first remote sensing image segment, and determining whether the network relationship link includes a network relationship link having a network connection line with the first remote sensing image segment, and determining whether the last remote sensing image segment included in the network relationship link belongs to the first remote sensing image segment, and then determining whether the first remote sensing image segment included in the network relationship link belongs to a second remote sensing image segment having a network connection line with the first remote sensing image segment;

[0039] If the network relationship link includes the first remote sensing image segment and includes a network relationship link with a network connection line between the first remote sensing image segment, the last remote sensing image segment included in the network relationship link belongs to the first remote sensing image segment, and whether the first remote sensing image segment included in the network relationship link belongs to the second remote sensing image segment with a network connection line between the first remote sensing image segment, then the network relationship link is marked as a target network relationship link.

[0040] In some preferred embodiments, in the above-mentioned remote sensing-based human settlement environment monitoring method, the step of performing correlation analysis on the output multiple target remote sensing image sets based on the included human settlement environment remote sensing images to output the target environment change analysis results of the target human settlement environment includes:

[0041] Calculating image similarity on the remote sensing image of the human settlement environment to output the image similarity;

[0042] The following steps are performed for each set of two target remote sensing images whose image acquisition times are adjacent:

[0043] performing one-to-one matching processing on the human settlement environment remote sensing images included in the two target remote sensing image sets based on the image similarity between every two frames of human settlement environment remote sensing images between the two target remote sensing image sets, and in accordance with a first matching principle of maximizing the mean value of the image similarities between the corresponding human settlement environment remote sensing images, and in accordance with a second matching principle of minimizing the mean value of the image time series differences between the corresponding human settlement environment remote sensing images, so as to form a one-to-one correspondence between the human settlement environment remote sensing images included in the two target remote sensing image sets;

[0044] fusing the image similarities between each pair of corresponding human settlement environment remote sensing images in the two target remote sensing image sets, and outputting the set correlation between the two target remote sensing image sets, wherein any pair of human settlement environment remote sensing images includes two frames of human settlement environment remote sensing images, and the two frames of human settlement environment remote sensing images belong to the two target remote sensing image sets respectively;

[0045] After obtaining the set correlation between each two target remote sensing image sets with adjacent image acquisition times, the target environment change analysis result of the target human settlement environment is determined based on the set correlation between each two adjacent target remote sensing image sets.

[0046] An embodiment of the present invention further provides a remote sensing-based human settlement environment monitoring system, which is applied to an environment monitoring server. The remote sensing-based human settlement environment monitoring system includes:

[0047] An image acquisition module is used to acquire images of a target human settlement environment and output a plurality of remote sensing image sets, each remote sensing image set including a plurality of remote sensing images of the human settlement environment, and the time interval between the image acquisition times corresponding to each two remote sensing image sets meets a preset time interval condition;

[0048] a remote sensing image screening module, configured to use an image processing neural network to screen the multiple human settlement environment remote sensing images included in each of the remote sensing image sets, so as to output a target remote sensing image set corresponding to each of the remote sensing image sets, wherein each of the target remote sensing image sets includes multiple human settlement environment remote sensing images;

[0049] The environmental change analysis module is used to perform correlation analysis on the output multiple target remote sensing image sets based on the included human settlement environment remote sensing images, so as to output the target environmental change analysis results of the target human settlement environment, and the target environmental change analysis results are used to reflect the degree of environmental change of the target human settlement environment.

[0050] The embodiment of the present invention provides a method and system for monitoring a human settlement environment based on remote sensing, which collects images of a target human settlement environment and outputs multiple remote sensing image sets. An image processing neural network is used to screen the multiple human settlement environment remote sensing images included in each remote sensing image set, so as to output a target remote sensing image set corresponding to each remote sensing image set, and each target remote sensing image set includes multiple human settlement environment remote sensing images. Based on the included human settlement environment remote sensing images, a correlation analysis is performed on the output multiple target remote sensing image sets to output a target environment change analysis result of the target human settlement environment. According to the foregoing content, since an image processing neural network is used to screen the multiple human settlement environment remote sensing images included in the remote sensing image set after the correlation analysis is performed, the accuracy of the target remote sensing image set is improved, so that the reliability of the target environment change analysis result of the target human settlement environment output based on the correlation analysis of the target remote sensing image set can be improved to a certain extent, thereby improving the reliability of human settlement environment monitoring to a certain extent.

[0051] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 This is a structural block diagram of the environment monitoring server provided by an embodiment of the present invention.

[0053] Figure 2 A flowchart of the steps of the method for monitoring a human settlement environment based on remote sensing provided in an embodiment of the present invention.

[0054] Figure 3 A schematic diagram of the modules included in the remote sensing-based human settlement environment monitoring system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings 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. The components of the embodiments of the present invention generally described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0056] Reference Figure 1 The embodiment of the present invention provides an environment monitoring server. The environment monitoring server may include a memory and a processor.

[0057] It should be noted that in some examples, the memory and the processor are electrically connected directly or indirectly to achieve data transmission or interaction. For example, they can be electrically connected to each other through one or more communication buses or signal lines. The memory may store at least one software function module (computer program) that can exist in the form of software or firmware. The processor can be used to execute the executable computer program stored in the memory, thereby implementing the remote sensing-based human settlement environment monitoring method provided in an embodiment of the present invention.

[0058] It should be noted that, in some examples, the memory may be, but is not limited to, a random access memory (RAM), a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), etc. The processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), a system on a chip (SoC), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0059] It should be noted that in some examples, Figure 1 The structure shown is only for illustration, and the environment monitoring server may also include Figure 1 More or fewer components than shown, or with Figure 1 The different configurations shown may, for example, include a communication unit for exchanging information with other devices.

[0060] Reference Figure 2 In addition to the content of the present invention, an embodiment of the present invention further provides a method for monitoring a human settlement environment based on remote sensing, which can be applied to the above-mentioned environment monitoring server. Among them, the method steps defined in the process related to the method for monitoring a human settlement environment based on remote sensing can be implemented by the environment monitoring server.

[0061] The following will Figure 2 The specific process shown is explained in detail.

[0062] Step S110 , collecting images of the target human settlement environment and outputting a plurality of remote sensing image sets.

[0063] In an embodiment of the present invention, the environmental monitoring server may capture images of a target human settlement environment and output multiple remote sensing image sets. Each remote sensing image set includes multiple remote sensing images of the human settlement environment, and the time interval between the image capture times of each two remote sensing image sets meets a preset time interval condition (e.g., an interval of one day, one week, one month, etc.).

[0064] Step S120 , using an image processing neural network to screen the multiple human settlement environment remote sensing images included in each of the remote sensing image sets, so as to output a target remote sensing image set corresponding to each of the remote sensing image sets.

[0065] In an embodiment of the present invention, the environment monitoring server may employ a (pre-trained) image processing neural network to screen the multiple human settlement environment remote sensing images included in each remote sensing image set, thereby outputting a target remote sensing image set corresponding to each remote sensing image set. Each target remote sensing image set includes multiple human settlement environment remote sensing images.

[0066] Step S130 , performing correlation analysis on the output multiple target remote sensing image sets based on the included human settlement environment remote sensing images, to output target environment change analysis results of the target human settlement environment.

[0067] In an embodiment of the present invention, the environment monitoring server may perform a correlation analysis on the outputted multiple target remote sensing image sets based on the included human settlement environment remote sensing images to output a target environment change analysis result of the target human settlement environment. The target environment change analysis result is used to reflect the degree of environmental change of the target human settlement environment (for example, the greater the correlation between two target remote sensing image sets that are adjacent in time sequence per image, the lower the degree of environmental change).

[0068] Through the above content, since after the correlation analysis, the image processing neural network will be used to screen the multiple human settlement environment remote sensing images included in the remote sensing image set, thereby improving the accuracy of the target remote sensing image set, the reliability of the target environmental change analysis results of the target human settlement environment output based on the correlation analysis of the target remote sensing image set can be improved to a certain extent, thereby improving the reliability of human settlement environment monitoring to a certain extent.

[0069] It should be noted that, in some examples, in the description corresponding to the human settlement environment monitoring method, step S110 may further include the following details:

[0070] After performing any image acquisition of the target human settlement environment, outputting a remote sensing image set corresponding to the image acquisition and determining acquisition parameters of the image acquisition;

[0071] Before performing the next image acquisition on the target human living environment, statistical processing of environmental events is performed on the target human living environment to output a statistical value of the environmental events currently corresponding to the target human living environment, wherein the environmental events are used to reflect events that have occurred in the time periods corresponding to two corresponding image acquisitions and have caused changes to the target human living environment (such as changes in personnel due to house sales, changes in buildings due to building construction and demolition, etc.), and the next image acquisition refers to an image acquisition performed after any of the above image acquisitions;

[0072] Adjusting acquisition parameters corresponding to any one image acquisition according to the environmental event statistics to output target acquisition parameters corresponding to the next image acquisition (for example, the target acquisition parameter may refer to the number of collected human settlement environment remote sensing images; the larger the environmental event statistics, the larger the target acquisition parameter may be);

[0073] According to the target acquisition parameters, the next image acquisition of the target human settlement environment is performed (through the corresponding acquisition equipment) to output a remote sensing image set corresponding to the image acquisition.

[0074] It should be noted that, in some examples, in the description corresponding to the human settlement environment monitoring method, step S120 may further include the following details:

[0075] Performing image segmentation on the remote sensing image set to form an image segmentation segment set, the image segmentation segment set including a first remote sensing image segment and a number greater than or equal to 1 second remote sensing image segment (wherein the first remote sensing image segment may be a remote sensing image segment having the largest average similarity with other remote sensing image segments in the image segmentation segment set, or the number of image objects included in the first remote sensing image segment has the largest number among the number of image objects included in each remote sensing image segment), the first remote sensing image segment including at least one remote sensing image of a human settlement environment, each of the second remote sensing image segments including at least one remote sensing image of a human settlement environment, when the first remote sensing image segment includes multiple remote sensing images of a human settlement environment, the multiple remote sensing images of the human settlement environment are continuous in image time sequence, and when the second remote sensing image segment includes multiple remote sensing images of a human settlement environment, the multiple remote sensing images of the human settlement environment are continuous in image time sequence;

[0076] Loading the first remote sensing image segments and the second remote sensing image segments greater than or equal to one into an image processing neural network, processing the first remote sensing image segments and the second remote sensing image segments greater than or equal to one using the image processing neural network, and outputting image segment close relationships and close relationship label information corresponding to each of the second remote sensing image segments greater than or equal to one, wherein each of the image segment close relationships is used to reflect whether the corresponding second remote sensing image segment and the first remote sensing image segment are matching image segments (for example, if an image object included in the second remote sensing image segment and an image object included in the first remote sensing image segment have a matching relationship in an object relationship database, then the image segment is determined to be a matching image segment, etc.);

[0077] splicing an image segment relationship network based on the image segment close relationship and the close relationship label information corresponding to each second remote sensing image segment, wherein the image segment close relationship reflects whether there is a network connection line between corresponding positions in the image segment relationship network, and the close relationship label information reflects a connection line feature of the network connection line (the connection line feature may reflect that the corresponding positions are connected by one or more network connection lines, such as the relationship between A and B includes: A is connected to B via a network connection line, or A is connected to C via a network connection line, and C is further connected to B via a network connection line, that is, A and B are connected via two network connection lines, or, in other examples, A and B are connected via three network connection lines, that is, A is connected to C via a network connection line, C is further connected to D via a network connection line, and D is further connected to B via a network connection line);

[0078] Wandering in the image segment relationship network (wherein a specific wandering method can be determined with reference to corresponding result conditions, such as ensuring that each network relationship link included in the network link set includes the first remote sensing image segment, and any second remote sensing image segment in each of the network relationship links has a network connection line with the first remote sensing image segment in the image segment relationship network), forming a network link set, and performing link screening on the network link set, outputting a target network relationship link corresponding to the image segmentation set, wherein each network relationship link included in the network link set includes the first remote sensing image segment, and any second remote sensing image segment in each of the network relationship links has a network connection line with the first remote sensing image segment in the image segment relationship network;

[0079] Based on the first remote sensing image segment and each second remote sensing image segment included in each target network relationship link, the target remote sensing image set corresponding to the remote sensing image set is screened out (for example, each human settlement remote sensing image included in the first remote sensing image segment and each second remote sensing image segment included in the target network relationship link can be assigned to the target remote sensing image set, or each human settlement remote sensing image included in the first remote sensing image segment can be assigned to the target remote sensing image set, and then the human settlement remote sensing image in each second remote sensing image segment whose image similarity with the human settlement remote sensing image included in the first remote sensing image segment is greater than or equal to a certain value is assigned to the target remote sensing image set).

[0080] It should be noted that, in some examples, in the description corresponding to the human settlement environment monitoring method, the step of loading the first remote sensing image segment and the second remote sensing image segment greater than or equal to 1 into an image processing neural network, processing the first remote sensing image segment and the second remote sensing image segment greater than or equal to 1 using the image processing neural network, and outputting the image segment close relationship and close relationship label information corresponding to each of the second remote sensing image segments greater than or equal to 1 may further include the following details:

[0081] Loading the first remote sensing image segment and the second remote sensing image segment whose number is greater than or equal to 1 into an image processing neural network, processing the first remote sensing image segment and the second remote sensing image segment whose number is greater than or equal to 1 by using the image processing neural network (the first remote sensing image segment can be represented by a low-dimensional vector by performing corresponding embedding processing, or encoding processing can be performed by a bidirectional encoding network to output corresponding vectors), outputting an image variable mapping continuous vector corresponding to the first remote sensing image segment, and outputting an image variable mapping continuous vector and an image meaning vector corresponding to each of the second remote sensing image segments whose number is greater than or equal to 1 (the image meaning vector is used to reflect the image information of the second remote sensing image segment);

[0082] According to the image time sequence relationship of the remote sensing image segments included in the image split segment set (the image time sequence of each remote sensing image segment is determined according to the image time sequence of the human settlement environment remote sensing image included in the remote sensing image segment, in this way, the first remote sensing image segment can be first combined with each other remote sensing image segment, and then the second remote sensing image segment can be combined with each other remote sensing image segment, and then the third remote sensing image segment can be combined with each other remote sensing image segment, and so on, until the last remote sensing image segment is combined with each other remote sensing image segment. After the combination is completed, the overlapping combinations are screened out, and the first remote sensing image segment and the second remote sensing image segment whose number is greater than or equal to 1 are randomly paired (such as the first remote sensing image segment), and the corresponding number of image segment paired combinations greater than or equal to 2 is output;

[0083] Using the image processing neural network, perform image vector splicing processing on the image segment pairing combinations greater than or equal to 2, and output image segment close relationships and close relationship label information corresponding to each image segment pairing combination;

[0084] Based on the image segment close relationships and close relationship label information corresponding to the image segment pairing combinations greater than or equal to 2, the image segment close relationships and close relationship label information corresponding to the second remote sensing image segments greater than or equal to 1 are determined.

[0085] It should be noted that, in some examples, in the description corresponding to the human settlement environment monitoring method, the step of using the image processing neural network to perform image vector splicing processing on the image segment pairing combinations greater than or equal to 2, and outputting the image segment close relationship and close relationship label information corresponding to each image segment pairing combination, may further include the following details:

[0086] Using an image vector splicing unit (which may be an artificial neural network, ANN) included in the image processing neural network, the image variable mapping continuous vector and the image meaning vector corresponding to the second remote sensing image segment included in each of the image segment pairing combinations are subjected to image vector splicing processing with the image variable mapping continuous vector corresponding to the first remote sensing image segment, and the image splicing vector corresponding to each of the image segment pairing combinations is output;

[0087] Utilizing the discriminant output function (which may be a softmax function or an activation function) included in the image processing neural network, vector discrimination is performed on each of the image splicing vectors, and the image segment close relationships and close relationship label information corresponding to the image segment pairing combinations whose number is greater than or equal to 2 are output.

[0088] It should be noted that, in some examples, in the description corresponding to the human settlement environment monitoring method, the step of determining the image segment close relationships and close relationship label information corresponding to the second remote sensing image segments greater than or equal to 1 based on the image segment close relationships and close relationship label information corresponding to the image segment pairing combinations greater than or equal to 2 may further include the following details:

[0089] Each of the image segment pairing combinations whose number is greater than or equal to 2 is processed by the following steps in sequence or simultaneously:

[0090] If the image segment pairing combination consists of a second remote sensing image segment, marking the image segment close relationship and close relationship label information corresponding to the image segment pairing combination as the image segment close relationship and close relationship label information corresponding to the second remote sensing image segment;

[0091] If the image segment pairing combination is composed of a plurality of second remote sensing image segments, the image segment close relationship and close relationship label information corresponding to the image segment pairing combination are marked as the image segment close relationship and close relationship label information corresponding to the second remote sensing image segment that is the earliest in image time sequence among the plurality of second remote sensing image segments;

[0092] Based on the image segment close relationship and close relationship label information corresponding to the one second remote sensing image segment, or based on the image segment close relationship and close relationship label information corresponding to the second remote sensing image segment at the front of the image time sequence, determine the image segment close relationship and close relationship label information corresponding to the second remote sensing image segments whose number is greater than or equal to 1.

[0093] It should be noted that, in some examples, in the description corresponding to the human settlement environment monitoring method, the step of splicing to form an image segment relationship network based on the image segment close relationship and the close relationship label information corresponding to each second remote sensing image segment may further include the following details:

[0094] Marking the network distribution information of the second remote sensing image segments and the first remote sensing image segments respectively having a number greater than or equal to 1, based on the image segment close relationship and the close relationship label information corresponding to each second remote sensing image segment;

[0095] Based on the network distribution information corresponding to each of the second remote sensing image segments and the network distribution information corresponding to the first remote sensing image segment, a connection direction with the second remote sensing image segment as the connection starting point and the first remote sensing image segment as the connection end point is used to splice the second remote sensing image segments and the first remote sensing image segments whose number is greater than or equal to 1 to form an image segment relationship network.

[0096] It should be noted that, in some examples, in the description corresponding to the human settlement environment monitoring method, the step of performing link screening on the network link set and outputting the target network relationship link corresponding to the image segmentation segment set may further include the following details:

[0097] Perform screening processing on each of the network relationship links included in the network link set by sequentially or synchronously performing the following steps:

[0098] Identifying the network relationship link to determine whether the network relationship link includes the first remote sensing image segment, and determining whether the network relationship link includes a network relationship link having a network connection line with the first remote sensing image segment, and determining whether the last remote sensing image segment included in the network relationship link belongs to the first remote sensing image segment, and then determining whether the first remote sensing image segment included in the network relationship link belongs to a second remote sensing image segment having a network connection line with the first remote sensing image segment;

[0099] If the network relationship link includes the first remote sensing image segment and includes a network relationship link with a network connection line between the first remote sensing image segment, the last remote sensing image segment included in the network relationship link belongs to the first remote sensing image segment, and whether the first remote sensing image segment included in the network relationship link belongs to the second remote sensing image segment with a network connection line between the first remote sensing image segment, then the network relationship link is marked as a target network relationship link.

[0100] It should be noted that, in some examples, in the description corresponding to the human settlement environment monitoring method, step S130 may further include the following details:

[0101] Calculating image similarity on the remote sensing image of the human settlement environment to output the image similarity;

[0102] The following steps are performed for each set of two target remote sensing images whose image acquisition times are adjacent:

[0103] performing one-to-one matching processing on the human settlement environment remote sensing images included in the two target remote sensing image sets based on the image similarity between every two frames of human settlement environment remote sensing images between the two target remote sensing image sets, and in accordance with a first matching principle of maximizing the mean value of the image similarities between the corresponding human settlement environment remote sensing images, and in accordance with a second matching principle of minimizing the mean value of the image time series differences between the corresponding human settlement environment remote sensing images, so as to form a one-to-one correspondence between the human settlement environment remote sensing images included in the two target remote sensing image sets;

[0104] fusing the image similarities between each pair of corresponding human settlement environment remote sensing images in the two target remote sensing image sets (e.g., performing mean calculation on the image similarities between each pair of corresponding human settlement environment remote sensing images), and outputting a set correlation between the two target remote sensing image sets, where any pair of human settlement environment remote sensing images includes two frames of human settlement environment remote sensing images, and the two frames of human settlement environment remote sensing images belong to the two target remote sensing image sets respectively;

[0105] After obtaining the set correlation between every two target remote sensing image sets with adjacent image acquisition times, the target environment change analysis result of the target human settlement environment is determined based on the set correlation between every two adjacent target remote sensing image sets (that is, the mean or product of the set correlation between every two adjacent target remote sensing image sets can be calculated, and then the target environment change analysis result is determined based on the calculation result; the smaller the calculation result, the greater the degree of environmental change reflected by the target environment change analysis result).

[0106] It should be noted that, in some examples, in the description corresponding to the human settlement environment monitoring method, the step of calculating image similarity of the human settlement environment remote sensing image to output the image similarity may further include the following details:

[0107] Performing pixel unit segmentation processing on the first human settlement environment remote sensing image to form a first pixel unit sequence corresponding to the first human settlement environment remote sensing image, where the first pixel unit sequence includes a plurality of first pixel units, each of the first pixel units includes at least one first pixel point, and for each first pixel unit including the plurality of first pixel points, a difference between pixel values of any two adjacent first pixel points in the first pixel unit is less than a preconfigured pixel difference threshold (that is, if the pixel value difference between adjacent first pixel points is small, they can be combined to form a first pixel unit; if the difference is large, they belong to different first pixel units);

[0108] performing pixel unit segmentation processing on the second human settlement environment remote sensing image to form a second pixel unit sequence corresponding to the second human settlement environment remote sensing image, where the second pixel unit sequence includes a plurality of second pixel units, each of the second pixel units includes at least one second pixel point, and for each second pixel unit including the plurality of second pixel points, a difference between pixel values of any two adjacent second pixel points in the second pixel unit is less than a pixel difference threshold (that is, if the difference in pixel values between adjacent second pixel points is small, they can be combined to form a single second pixel unit; if the difference is large, they belong to different second pixel units);

[0109] For each first pixel unit, the pixel values of at least one first pixel point included in the first pixel unit are fused (such as by calculating the mean value), and a representative pixel value corresponding to the first pixel unit is output; for each second pixel unit, the pixel values of at least one second pixel point included in the second pixel unit are fused, and a representative pixel value corresponding to the second pixel unit is output;

[0110] Extracting each target first pixel unit from the first pixel unit sequence based on the representative pixel value corresponding to each first pixel unit, and then extracting each target second pixel unit from the second pixel unit sequence based on the representative pixel value corresponding to each second pixel unit, wherein the difference between the representative pixel value corresponding to each target first pixel unit and the representative pixel value corresponding to each other first pixel unit within an adjacent sequence position range (the specific size is not limited, for example, it can be within 5, 6, or 10 pixels in the sequence) is less than a preconfigured pixel difference reference value, and the difference between the representative pixel value corresponding to each target second pixel unit and the representative pixel value corresponding to each other second pixel unit within the adjacent sequence position range is less than the pixel difference reference value;

[0111] Extracting image feature points from the first remote sensing image of the human settlement environment (referring to feature point extraction methods in the prior art, such as the Oriented FAST and Rotated BRIEF algorithms) to output a first feature point set corresponding to the first remote sensing image of the human settlement environment, and then extracting image feature points from the second remote sensing image of the human settlement environment to output a second feature point set corresponding to the second remote sensing image of the human settlement environment, where the first feature point set includes a plurality of first image feature points, and the second feature point set includes a plurality of second image feature points;

[0112] For each of the first image feature points, based on the pixel value of each first pixel point in the pixel area where the first image feature point is located (for example, a pixel area can be determined with the first image feature point as the center point and a certain size and shape), the pixel feature corresponding to the first image feature point is determined (the pixel feature can be represented by a feature matrix, and the feature matrix can include two rows, one of which represents the pixel position relationship between each first pixel point and the first image feature point, and the other row represents the pixel difference between each first pixel point and the first image feature point). For each of the second image feature points, based on the pixel value of each second pixel point in the pixel area where the second image feature point is located the pixel value of the point, determining the pixel feature corresponding to the second image feature point, and then performing a one-to-one matching process on the first image feature points included in the first feature point set and the second image feature points included in the second feature point set based on the pixel feature corresponding to each of the first image feature points and the pixel feature corresponding to each of the second image feature points (on the principle of maximizing the feature similarity between them) (i.e., marking the first image feature points with the matching second image feature points as target first image feature points), so as to mark each target first image feature point from the first feature point set, and mark the target second image feature point corresponding to each target first image feature point from the second feature point set;

[0113] Screening out each target first pixel unit including at least one target first image feature point to mark it as an important target first pixel unit, then screening out each target second pixel unit including at least one target second image feature point to mark it as an important target second pixel unit, and comparing and analyzing a representative pixel position (the representative pixel position may be a position of an intermediate pixel point of the important target first pixel unit) and a representative pixel value corresponding to each of the important target first pixel units with a representative pixel position and a representative pixel value corresponding to each of the important target second pixel units to determine a comprehensive difference coefficient (i.e., a comprehensive difference in pixel position and pixel value) between the important target first pixel unit and the important target second pixel unit;

[0114] Based on the comprehensive difference coefficient, the proportion of the number of the target first image feature points in the first feature point set or the proportion of the number of the target second image feature points in the second feature point set is adjusted, and the image similarity between the first human settlement environment remote sensing image and the second human settlement environment remote sensing image is output (for example, the image similarity can be inversely proportional to the comprehensive difference coefficient and directly proportional to the proportion).

[0115] Reference Figure 3In addition to the content of the present invention, an embodiment of the present invention further provides a remote sensing-based human settlement environment monitoring system that can be applied to the above-mentioned environment monitoring server. The remote sensing-based human settlement environment monitoring system may include the following software functional modules, such as an image acquisition module, a remote sensing image screening module, and an environmental change analysis module, or other software functional modules.

[0116] It should be noted that in some examples, the image acquisition module is used to capture images of the target human settlement environment and output multiple remote sensing image sets, each remote sensing image set includes multiple remote sensing images of the human settlement environment, and the time interval between the image acquisition times corresponding to each two remote sensing image sets meets the preset time interval condition.

[0117] It should be noted that in some examples, the remote sensing image screening module is used to use an image processing neural network to screen the multiple human settlement environment remote sensing images included in each of the remote sensing image sets, so as to output a target remote sensing image set corresponding to each of the remote sensing image sets, and each of the target remote sensing image sets includes multiple human settlement environment remote sensing images.

[0118] It should be noted that in some examples, the environmental change analysis module is used to perform correlation analysis on the output multiple target remote sensing image sets based on the included human settlement environment remote sensing images, so as to output the target environmental change analysis results of the target human settlement environment, and the target environmental change analysis results are used to reflect the degree of environmental change of the target human settlement environment.

[0119] In summary, the present invention provides a method and system for monitoring a human settlement environment based on remote sensing, which collects images of a target human settlement environment and outputs multiple remote sensing image sets. An image processing neural network is used to screen the multiple human settlement environment remote sensing images included in each remote sensing image set, so as to output a target remote sensing image set corresponding to each remote sensing image set, and each target remote sensing image set includes multiple human settlement environment remote sensing images. Based on the included human settlement environment remote sensing images, a correlation analysis is performed on the output multiple target remote sensing image sets to output a target environment change analysis result of the target human settlement environment. Through the foregoing content, since after the correlation analysis is performed, an image processing neural network is used to screen the multiple human settlement environment remote sensing images included in the remote sensing image set, thereby improving the accuracy of the target remote sensing image set, so that the reliability of the target environment change analysis result of the target human settlement environment output based on the correlation analysis of the target remote sensing image set can be improved to a certain extent, thereby improving the reliability of human settlement environment monitoring to a certain extent.

[0120] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A method for monitoring human settlement environment based on remote sensing, characterized in that: Applied to an environmental monitoring server, the remote sensing-based human settlement environment monitoring method includes: Capturing images of a target human settlement environment and outputting a plurality of remote sensing image sets, each of the remote sensing image sets comprising a plurality of remote sensing images of the human settlement environment, wherein the time interval between image acquisition times corresponding to each two remote sensing image sets satisfies a preset time interval condition; Using an image processing neural network, screening the multiple human settlement environment remote sensing images included in each of the remote sensing image sets to output a target remote sensing image set corresponding to each of the remote sensing image sets, each of the target remote sensing image sets including the multiple human settlement environment remote sensing images; Based on the included human settlement environment remote sensing images, correlation analysis is performed on the output multiple target remote sensing image sets to output a target environment change analysis result of the target human settlement environment, wherein the target environment change analysis result is used to reflect the degree of environmental change of the target human settlement environment; The step of using an image processing neural network to screen the multiple human settlement environment remote sensing images included in each remote sensing image set to output a target remote sensing image set corresponding to each remote sensing image set includes: Performing image segmentation on the remote sensing image set to form an image segmentation segment set, the image segmentation segment set including a first remote sensing image segment and a number greater than or equal to one second remote sensing image segment, the first remote sensing image segment including at least one human settlement remote sensing image, each of the second remote sensing image segments including at least one human settlement remote sensing image, when the first remote sensing image segment includes a plurality of human settlement remote sensing images, the plurality of human settlement remote sensing images are continuous in image time sequence, and when the second remote sensing image segment includes a plurality of human settlement remote sensing images, the plurality of human settlement remote sensing images are continuous in image time sequence; Loading the first remote sensing image segments and the second remote sensing image segments greater than or equal to one into an image processing neural network, processing the first remote sensing image segments and the second remote sensing image segments greater than or equal to one using the image processing neural network, and outputting image segment close relationships and close relationship label information corresponding to each of the second remote sensing image segments greater than or equal to one, wherein each image segment close relationship is used to reflect whether the corresponding second remote sensing image segment and the first remote sensing image segment are matching image segments; splicing an image segment relationship network based on the image segment close relationship and the close relationship label information corresponding to each second remote sensing image segment, wherein the image segment close relationship reflects whether there is a network connection line between corresponding positions in the image segment relationship network, and the close relationship label information reflects the connection line characteristics of the network connection line; Wandering in the image segment relationship network to form a network link set, and performing link screening on the network link set to output target network relationship links corresponding to the image segmentation segment set, wherein each network relationship link included in the network link set includes the first remote sensing image segment, and any second remote sensing image segment in each network relationship link has a network connection line with the first remote sensing image segment in the image segment relationship network; According to the first remote sensing image segment and each second remote sensing image segment included in each target network relationship link, a target remote sensing image set corresponding to the remote sensing image set is screened out.

2. The method for monitoring human settlement environment based on remote sensing according to claim 1, wherein: The step of collecting images of the target human settlement environment and outputting a plurality of remote sensing image sets includes: After performing any image acquisition of the target human settlement environment, outputting a remote sensing image set corresponding to the image acquisition and determining acquisition parameters of the image acquisition; Before performing a next image acquisition on the target human living environment, statistical processing of environmental events is performed on the target human living environment to output a statistical value of environmental events currently corresponding to the target human living environment, wherein the environmental events are used to reflect events that occur in time periods corresponding to two corresponding image acquisitions and cause changes to the target human living environment, and the next image acquisition refers to an image acquisition performed after any of the image acquisitions; Adjusting acquisition parameters corresponding to any one image acquisition according to the environmental event statistics to output target acquisition parameters corresponding to the next image acquisition; The next image acquisition is performed on the target human settlement environment according to the target acquisition parameters to output a remote sensing image set corresponding to the image acquisition.

3. The method for monitoring human settlement environment based on remote sensing according to claim 1, wherein: The step of loading the first remote sensing image segment and the second remote sensing image segment greater than or equal to 1 into an image processing neural network, processing the first remote sensing image segment and the second remote sensing image segment greater than or equal to 1 using the image processing neural network, and outputting image segment close relationships and close relationship label information corresponding to each of the second remote sensing image segments greater than or equal to 1 includes: Loading the first remote sensing image segment and the second remote sensing image segment whose number is greater than or equal to 1 into an image processing neural network, processing the first remote sensing image segment and the second remote sensing image segment whose number is greater than or equal to 1 using the image processing neural network, outputting an image variable mapping continuous vector corresponding to the first remote sensing image segment, and outputting an image variable mapping continuous vector and an image meaning vector corresponding to each of the second remote sensing image segments whose number is greater than or equal to 1; performing random pairing processing on the first remote sensing image segment and the second remote sensing image segment greater than or equal to 1 according to an image time sequence relationship of the remote sensing image segments included in the image segmentation segment set, and outputting corresponding image segment pairing combinations greater than or equal to 2; Using the image processing neural network, perform image vector splicing processing on the image segment pairing combinations greater than or equal to 2, and output image segment close relationships and close relationship label information corresponding to each image segment pairing combination; Based on the image segment close relationships and close relationship label information corresponding to the image segment pairing combinations greater than or equal to 2, the image segment close relationships and close relationship label information corresponding to the second remote sensing image segments greater than or equal to 1 are determined.

4. The method for monitoring human settlement environment based on remote sensing according to claim 3, wherein: The step of performing image vector splicing processing on the image segment pairing combinations greater than or equal to 2 using the image processing neural network and outputting the image segment close relationship and close relationship label information corresponding to each image segment pairing combination includes: Using an image vector splicing unit included in the image processing neural network, the image variable mapping continuous vector and the image meaning vector corresponding to the second remote sensing image segment included in each of the image segment pairing combinations are subjected to image vector splicing processing with the image variable mapping continuous vector corresponding to the first remote sensing image segment, and the image splicing vector corresponding to each of the image segment pairing combinations is output; Using the discriminant output function included in the image processing neural network, vector discrimination is performed on each of the image splicing vectors, and the image segment close relationships and close relationship label information corresponding to the image segment pairing combinations greater than or equal to 2 are output.

5. The method for monitoring human settlement environment based on remote sensing according to claim 3, wherein: The step of determining the image segment close relationships and close relationship label information corresponding to the second remote sensing image segments greater than or equal to 1 based on the image segment close relationships and close relationship label information corresponding to the image segment pairing combinations greater than or equal to 2, comprises: Each of the image segment pairing combinations whose number is greater than or equal to 2 is processed by the following steps in sequence or simultaneously: If the image segment pairing combination consists of a second remote sensing image segment, marking the image segment close relationship and close relationship label information corresponding to the image segment pairing combination as the image segment close relationship and close relationship label information corresponding to the second remote sensing image segment; If the image segment pairing combination is composed of a plurality of second remote sensing image segments, the image segment close relationship and close relationship label information corresponding to the image segment pairing combination are marked as the image segment close relationship and close relationship label information corresponding to the second remote sensing image segment that is the earliest in image time sequence among the plurality of second remote sensing image segments; Based on the image segment close relationship and close relationship label information corresponding to the one second remote sensing image segment, or based on the image segment close relationship and close relationship label information corresponding to the second remote sensing image segment at the front of the image time sequence, determine the image segment close relationship and close relationship label information corresponding to the second remote sensing image segments whose number is greater than or equal to 1.

6. The method for monitoring human settlement environment based on remote sensing according to claim 1, wherein: The step of forming an image segment relationship network by stitching together the image segment close relationship and the close relationship label information corresponding to each second remote sensing image segment includes: Marking the network distribution information of the second remote sensing image segments and the first remote sensing image segments respectively having a number greater than or equal to 1, based on the image segment close relationship and the close relationship label information corresponding to each second remote sensing image segment; Based on the network distribution information corresponding to each of the second remote sensing image segments and the network distribution information corresponding to the first remote sensing image segment, a connection direction with the second remote sensing image segment as the connection starting point and the first remote sensing image segment as the connection end point is used to splice the second remote sensing image segments and the first remote sensing image segments whose number is greater than or equal to 1 to form an image segment relationship network.

7. The method for monitoring human settlement environment based on remote sensing according to claim 1, wherein: The step of screening the network link set and outputting the target network relationship links corresponding to the image segmentation segment set includes: Perform screening processing on each of the network relationship links included in the network link set by sequentially or synchronously performing the following steps: Identifying the network relationship link to determine whether the network relationship link includes the first remote sensing image segment, and determining whether the network relationship link includes a network relationship link having a network connection line with the first remote sensing image segment, and determining whether the last remote sensing image segment included in the network relationship link belongs to the first remote sensing image segment, and then determining whether the first remote sensing image segment included in the network relationship link belongs to a second remote sensing image segment having a network connection line with the first remote sensing image segment; If the network relationship link includes the first remote sensing image segment and includes a network relationship link with a network connection line between the first remote sensing image segment, the last remote sensing image segment included in the network relationship link belongs to the first remote sensing image segment, and whether the first remote sensing image segment included in the network relationship link belongs to the second remote sensing image segment with a network connection line between the first remote sensing image segment, then the network relationship link is marked as a target network relationship link.

8. The method for monitoring a human settlement environment based on remote sensing according to any one of claims 1 to 7, wherein: The step of performing correlation analysis on the output multiple target remote sensing image sets based on the included human settlement environment remote sensing images to output the target environment change analysis results of the target human settlement environment includes: Calculating image similarity on the remote sensing image of the human settlement environment to output the image similarity; The following steps are performed for each set of two target remote sensing images whose image acquisition times are adjacent: performing one-to-one matching processing on the human settlement environment remote sensing images included in the two target remote sensing image sets based on the image similarity between every two frames of human settlement environment remote sensing images between the two target remote sensing image sets, and in accordance with a first matching principle of maximizing the mean value of the image similarities between the corresponding human settlement environment remote sensing images, and in accordance with a second matching principle of minimizing the mean value of the image time series differences between the corresponding human settlement environment remote sensing images, so as to form a one-to-one correspondence between the human settlement environment remote sensing images included in the two target remote sensing image sets; fusing the image similarities between each pair of corresponding human settlement environment remote sensing images in the two target remote sensing image sets, and outputting the set correlation between the two target remote sensing image sets, wherein any pair of human settlement environment remote sensing images includes two frames of human settlement environment remote sensing images, and the two frames of human settlement environment remote sensing images belong to the two target remote sensing image sets respectively; After obtaining the set correlation between each two target remote sensing image sets with adjacent image acquisition times, the target environment change analysis result of the target human settlement environment is determined based on the set correlation between each two adjacent target remote sensing image sets.

9. A human settlement environment monitoring system based on remote sensing, characterized in that: Applied to an environmental monitoring server, the remote sensing-based human settlement environment monitoring system includes: An image acquisition module is used to acquire images of a target human settlement environment and output a plurality of remote sensing image sets, each remote sensing image set including a plurality of remote sensing images of the human settlement environment, and the time interval between the image acquisition times corresponding to each two remote sensing image sets meets a preset time interval condition; a remote sensing image screening module, configured to use an image processing neural network to screen the multiple human settlement environment remote sensing images included in each of the remote sensing image sets, so as to output a target remote sensing image set corresponding to each of the remote sensing image sets, wherein each of the target remote sensing image sets includes multiple human settlement environment remote sensing images; An environmental change analysis module is used to perform a correlation analysis on a plurality of output target remote sensing image sets based on the included human settlement environment remote sensing images, so as to output a target environmental change analysis result of the target human settlement environment, wherein the target environmental change analysis result is used to reflect the degree of environmental change of the target human settlement environment; The image processing neural network is used to screen the multiple human settlement environment remote sensing images included in each remote sensing image set to output a target remote sensing image set corresponding to each remote sensing image set, including: Performing image segmentation on the remote sensing image set to form an image segmentation segment set, the image segmentation segment set including a first remote sensing image segment and a number greater than or equal to one second remote sensing image segment, the first remote sensing image segment including at least one human settlement remote sensing image, each of the second remote sensing image segments including at least one human settlement remote sensing image, when the first remote sensing image segment includes a plurality of human settlement remote sensing images, the plurality of human settlement remote sensing images are continuous in image time sequence, and when the second remote sensing image segment includes a plurality of human settlement remote sensing images, the plurality of human settlement remote sensing images are continuous in image time sequence; Loading the first remote sensing image segments and the second remote sensing image segments greater than or equal to one into an image processing neural network, processing the first remote sensing image segments and the second remote sensing image segments greater than or equal to one using the image processing neural network, and outputting image segment close relationships and close relationship label information corresponding to each of the second remote sensing image segments greater than or equal to one, wherein each image segment close relationship is used to reflect whether the corresponding second remote sensing image segment and the first remote sensing image segment are matching image segments; splicing an image segment relationship network based on the image segment close relationship and the close relationship label information corresponding to each second remote sensing image segment, wherein the image segment close relationship reflects whether there is a network connection line between corresponding positions in the image segment relationship network, and the close relationship label information reflects the connection line characteristics of the network connection line; Wandering in the image segment relationship network to form a network link set, and performing link screening on the network link set to output target network relationship links corresponding to the image segmentation segment set, wherein each network relationship link included in the network link set includes the first remote sensing image segment, and any second remote sensing image segment in each network relationship link has a network connection line with the first remote sensing image segment in the image segment relationship network; According to the first remote sensing image segment and each second remote sensing image segment included in each target network relationship link, a target remote sensing image set corresponding to the remote sensing image set is screened out.

Images