Method, device, equipment, medium and product for monitoring remote sensing change of human activities in land natural reserve

Abstract

The invention discloses a method, a device, equipment, a medium and a product for monitoring remote sensing change of human activities in a land natural reserve, and relates to the technical field of remote sensing. The method comprises the following steps: respectively inputting two remote sensing images of an object to be monitored in different time phases into a super-resolution network to obtain two super-resolution images in different time phases; a first DenseNet network and a second DenseNet network are used for carrying out feature extraction on the two super-resolution images respectively, and two feature images of different time phases are obtained; the first double-attention module and the second double-attention module are used for conducting feature correction on the two feature maps respectively, and feature vectors obtained after different time phases are corrected are obtained; the two double-attention modules both adopt a space attention mechanism and a channel attention mechanism; and performing similarity analysis on the corrected feature vectors of the different time phases to obtain a change region of the to-be-monitored object between the two remote sensing images of the different time phases. According to the invention, the accuracy and efficiency of remote sensing image change monitoring can be improved.

Description

Technical Field

[0001] The present application relates to the field of remote sensing technology, and in particular to a method, device, equipment, medium and product for remote sensing change monitoring of human activities in terrestrial nature reserves. Background Art

[0002] In the process of monitoring changes in human activities in terrestrial nature reserves such as forests, grasslands, wetlands, houses, roads, and mining land, high-resolution remote sensing images have the advantage of high spatial resolution and can clearly distinguish the details of objects, so they have become an important means of identifying objects and monitoring changes. The method is to quickly interpret the changes between the two images by comparing the remote sensing images before and after at different scales such as provinces, cities, and counties. However, there are several significant problems with monitoring changes in objects using high-resolution remote sensing images:

[0003] First, the clarity of high-resolution remote sensing images is not very stable. For objects with small changes in the area of ​​the image (such as houses, roads, and other construction land, etc.), they are often not recognized, resulting in insufficient information when monitoring subtle changes in objects, which affects the accuracy of judgment. Second, when dealing with complex scenes or multi-source images, existing change monitoring algorithms are easily disturbed by noise and changes in illumination, resulting in false alarms and missed alarms. Third, the computational efficiency of the algorithm is also a major problem. Many high-precision algorithms are difficult to meet the needs of real-time monitoring, and the accuracy is not high. Summary of the invention

[0004] The purpose of this application is to provide a method, device, equipment, medium and product for remote sensing change monitoring of human activities in terrestrial nature reserves, which can improve the accuracy and efficiency of remote sensing image change monitoring.

[0005] To achieve the above objectives, this application provides the following solutions:

[0006] In the first aspect, the present application provides a method for remote sensing change monitoring of human activities in terrestrial nature reserves, including: inputting two remote sensing images of the object to be monitored at different phases into a super-resolution network respectively, to obtain two super-resolution images of different phases; the object to be monitored is human activities in a terrestrial nature reserve; using a first DenseNet network to extract features from a super-resolution image, and using a second DenseNet network to extract features from another super-resolution image, to obtain two feature images of different phases; using a first dual attention module to perform feature correction on one feature image, and using a second dual attention module to perform feature correction on another feature image, to obtain feature vectors after correction at different phases; both the first dual attention module and the second dual attention module adopt a spatial attention mechanism and a channel attention mechanism; performing similarity analysis on the corrected feature vectors of different phases to obtain the change area between the two remote sensing images of the object to be monitored at different phases.

[0007] In the second aspect, the present application provides a remote sensing change monitoring device for human activities in terrestrial nature reserves, including: a resolution conversion module, a feature extraction module, a feature correction module and a similarity analysis module. The resolution conversion module is used to input two remote sensing images of the monitored object in different phases into a super-resolution network respectively to obtain two super-resolution images of different phases; the monitored object is human activities in terrestrial nature reserves; the feature extraction module is used to use the first DenseNet network to extract features from a super-resolution image, and use the second DenseNet network to extract features from another super-resolution image, to obtain two feature images of different phases; the feature correction module is used to use the first dual attention module to perform feature correction on a feature image, and use the second dual attention module to perform feature correction on another feature image, to obtain feature vectors corrected in different phases; the first dual attention module and the second dual attention module both use spatial attention mechanism and channel attention mechanism; the similarity analysis module is used to perform similarity analysis on the feature vectors corrected in different phases to obtain the change area between the two remote sensing images of the monitored object in different phases.

[0008] In a third aspect, the present application provides a computer device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the above-described methods for remote sensing change monitoring of human activities in terrestrial nature reserves.

[0009] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for monitoring changes in human activities in terrestrial nature reserves through remote sensing.

[0010] In a fifth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned methods for monitoring changes in human activities in terrestrial nature reserves through remote sensing.

[0011] According to the specific embodiments provided in this application, this application has the following technical effects:

[0012] The present application provides a method, device, equipment, medium and product for monitoring changes in human activities in terrestrial nature reserves through remote sensing, which converts remote sensing images into higher-resolution images through a super-resolution network, improves the clarity and detail of the image, overcomes the defects of the prior art in monitoring subtle changes, which affects the accuracy due to insufficient information, and improves the spatial resolution of the image and the clarity of the object to be monitored; compared with the traditional change monitoring method, the DenseNet network can more accurately capture the changes of the target object under complex background changes and slight differences in target features, and provide more reliable monitoring results; and by introducing the dual attention mechanism of the spatial attention mechanism and the channel attention mechanism, it can significantly improve the model's ability to recognize important features of the target, thereby improving the accuracy of remote sensing image change monitoring. At the same time, the present application utilizes the recognition ability of the network to overcome the defect that the existing high-precision algorithm cannot meet the real-time monitoring needs, and improves the efficiency of remote sensing image change monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0014] Figure 1 This is an application environment diagram of a method for monitoring changes in human activities in a terrestrial nature reserve using remote sensing in an embodiment of the present application;

[0015] Figure 2 A schematic diagram of a flow chart of a method for monitoring changes in human activities in a terrestrial nature reserve using remote sensing provided in one embodiment of the present application;

[0016] Figure 3 A schematic diagram of the principle of a method for monitoring changes in human activities in a terrestrial nature reserve using remote sensing provided in one embodiment of the present application;

[0017] Figure 4 This is a schematic diagram of the SRNet super-resolution network architecture;

[0018] Figure 5A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0020] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0021] The method for monitoring changes in human activities in terrestrial nature reserves provided by the embodiment of the present application can be applied to Figure 1In the application environment shown, the terminal 102 communicates with the server 104 through a network. The data storage system can store data that the server 104 needs to process. The data storage system can be set up separately, integrated on the server 104, or placed on the cloud or other servers. The terminal 102 can send two remote sensing images of the object to be monitored at different phases to the server 104. After the server 104 receives the two remote sensing images of the object to be monitored at different phases, for the two remote sensing images of the object to be monitored at different phases, the server 104 inputs the two remote sensing images of the object to be monitored at different phases into a super-resolution network respectively to obtain two super-resolution images at different phases; the object to be monitored is human activities in a terrestrial nature reserve; a first DenseNet network is used to extract features from a super-resolution image, and a second DenseNet network is used to extract features from another super-resolution image to obtain two feature images at different phases; a first dual attention module is used to correct features of a feature image, and a second dual attention module is used to correct features of another feature image to obtain feature vectors after correction at different phases; the first dual attention module and the second dual attention module both adopt a spatial attention mechanism and a channel attention mechanism; similarity analysis is performed on the corrected feature vectors at different phases to obtain a change area between the two remote sensing images of the object to be monitored at different phases. The server 104 can feed back the obtained change area between the two remote sensing images of the object to be monitored at different time phases to the terminal 102. In addition, in some embodiments, the remote sensing change monitoring method for human activities in terrestrial nature reserves can also be implemented by the server 104 or the terminal 102 alone, such as the terminal 102 can directly monitor and process the two remote sensing images of the object to be monitored at different time phases, or the server 104 can obtain the two remote sensing images of the object to be monitored at different time phases from the data storage system, and monitor and process the two remote sensing images of the object to be monitored at different time phases.

[0022] The terminal 102 may be, but is not limited to, various desktop computers, laptop computers, smart phones, tablet computers, IoT devices, and portable wearable devices. The IoT devices may be smart speakers, smart TVs, smart air conditioners, smart vehicle-mounted devices, etc. The portable wearable devices may be smart watches, smart bracelets, head-mounted devices, etc. The server 104 may be implemented as an independent server or a server cluster consisting of multiple servers, or may be a cloud server.

[0023] In an exemplary embodiment, Figure 2As shown, a method for remote sensing changes in human activities in terrestrial nature reserves is provided. The method is executed by a computer device, and can be executed by a computer device such as a terminal or a server alone, or by a terminal and a server together. In the embodiment of the present application, the method is applied to Figure 1 The server 104 in the example is used as an example to illustrate the method, which includes the following steps 201 to 204. Among them:

[0024] Step 201: Input two remote sensing images of the object to be monitored at different time phases into a super-resolution network respectively to obtain two super-resolution images at different time phases. The object to be monitored is human activities in a terrestrial nature reserve.

[0025] Step 202: Use the first DenseNet network to perform feature extraction on one super-resolution image, and use the second DenseNet network to perform feature extraction on another super-resolution image to obtain two feature images at different phases.

[0026] Step 203: Use the first dual attention module to perform feature correction on a feature map, and use the second dual attention module to perform feature correction on another feature map to obtain feature vectors after correction in different time phases; both the first dual attention module and the second dual attention module adopt spatial attention mechanism and channel attention mechanism.

[0027] Step 204: Perform similarity analysis on the corrected feature vectors in different time phases to obtain the change region of the monitored object between the two remote sensing images in different time phases.

[0028] Implementing the above steps 201 to 204, the remote sensing image is converted into a higher resolution image through a super-resolution network, which improves the clarity and detail of the image, overcomes the defects of the prior art in monitoring subtle changes, which affects the accuracy due to insufficient information, and improves the spatial resolution of the image and the clarity of the monitored object; compared with the traditional change monitoring method, the DenseNet network can more accurately capture the changes of the target object under complex background changes and slight differences in target features, and provide more reliable monitoring results; and by introducing the dual attention mechanism of the spatial attention mechanism and the channel attention mechanism, it can significantly improve the model's ability to recognize important features of the target, thereby improving the accuracy of remote sensing image change monitoring. At the same time, this application utilizes the recognition ability of the network to overcome the defect that the existing high-precision algorithm cannot meet the real-time monitoring needs, and improves the efficiency of remote sensing image change monitoring.

[0029] In another exemplary embodiment of the present application, the super-resolution network (SRNet) in the above step 201 is intended to solve the blurring problem of low-resolution images in details and contours. In traditional image processing, low-resolution images often cannot clearly express the subtle features of objects, which has a negative impact on subsequent image analysis and understanding. By introducing super-resolution technology, we can convert low-resolution images into higher-resolution images, thereby improving the clarity and detail of the image. The super-resolution network generates high-quality high-resolution images using deep learning technology by learning a large number of high- and low-resolution image pairs. Compared with traditional interpolation methods, SRNet can better reconstruct edges and details in images, avoiding blur and artifacts. More importantly, the super-resolution network can effectively capture the complex structures in the image, enhance the visual effect of the image, and meet the higher standard application requirements of remote sensing image change monitoring.

[0030] Figure 3 In , we first obtain the two input images that need to be compared - input Figure 1 and input Figure 2 These images are usually remote sensing images from different time points, and may have insufficient clarity due to different shooting equipment, environmental conditions or image resolution. To overcome this problem, the two images are passed to the super-resolution network SRNet. SRNet is designed based on deep learning algorithms and aims to improve the quality of low-resolution images through reconstruction technology. The network generates super-resolution images by analyzing the texture, edges and details of the input image. Figure 1 and super-resolution Figure 2 . After super-resolution processing, the clarity and details of the image are significantly improved, which lays a solid foundation for subsequent feature extraction. Clear images not only help improve the accuracy of change monitoring, but also reduce misjudgments caused by blur. In addition, the results of super-resolution processing can effectively supplement the key information missing in the image due to low resolution, and improve the applicability of the image in different application scenarios. The optimization of this process provides a richer source of information for subsequent feature extraction, ensuring that the model can obtain more comprehensive and accurate data when analyzing images.

[0031] Figure 4 Figure 1 shows the SRNet super-resolution network architecture. The SRNet network extracts features from the input low-resolution image. These features usually include information such as the texture and edges of the image. In order to improve the expressiveness of the features, the SRNet network usually uses a deep convolutional neural network for feature extraction. Three dense blocks map the features of the low-resolution image to the high-resolution image space, and after upsampling and two convolutional layers, a high-resolution image is output.

[0032] For example, input Figure 1 and input Figure 2 It can be a high-resolution image, referred to as a high-resolution image.

[0033] In another exemplary embodiment of the present application, the super-resolution image obtained after super-resolution processing in step 202 is Figure 1 and super-resolution Figure 2 Input to the DenseNet network for feature extraction. DenseNet is a deep convolutional neural network that enhances the transfer and reuse of features by connecting each layer to all previous layers. In the feature extraction process, a shared weight mechanism is adopted, which means that super-resolution Figure 1 and super-resolution Figure 2 Use the same network weights for feature learning. This design not only effectively reduces the number of model parameters and improves training efficiency, but also ensures the consistency of the feature learning process and enhances the accuracy of feature comparison between two images. In this way, the model can better capture the subtle differences and similarities between images, and effectively identify the changed areas. In addition, the structural characteristics of DenseNet make the network extract richer features at different levels, providing multi-dimensional information support for subsequent analysis. The quality of feature extraction directly affects the effect of remote sensing change monitoring. Therefore, in this link, ensuring that the extracted features are both distinguishable and representative is the key to achieving efficient remote sensing change monitoring.

[0034] The weights are shared during the extraction process of the two images, and DenseNet is used for feature extraction, which allows efficient reuse and transfer of features, improves the consistency of feature learning, and ensures accurate capture of subtle differences between images.

[0035] In another exemplary embodiment of the present application, the above step 203 uses a dual attention module to perform feature correction on the feature map, and the dual attention module adopts a spatial attention mechanism and a channel attention mechanism. Attention mechanism plays a vital role in deep learning, especially in image and natural language processing tasks. By introducing a dual attention mechanism of a spatial attention mechanism and a channel attention mechanism, the model's ability to recognize important features of the target can be significantly improved. The spatial attention mechanism focuses on specific areas in the image to ensure that the model can focus on those parts that contribute significantly to the task during processing, thereby improving the accuracy of feature extraction. The channel attention mechanism dynamically adjusts the weights of different channels to suppress redundant features and ensure that the model is more focused on representative features. Compared with traditional methods, the dual attention mechanism can effectively reduce feature interference and enhance the model's sensitivity to target features, thereby improving the performance of tasks such as image classification and target monitoring. This mechanism not only improves the performance of the model, but also provides a visual basis for understanding the decision-making process of the model, making the final result more interpretable.

[0036] In another exemplary embodiment of the present application, the first dual attention module and the second dual attention module both include: a spatial attention module and a channel attention module. After introducing the spatial attention and channel attention modules, the model can focus on key information areas, dynamically adjust feature weights, and significantly improve sensitivity and accuracy to changes.

[0037] In another exemplary embodiment of the present application, the above step 203 uses the first dual attention module to perform feature correction on a feature map, which can be replaced by the following steps 301 to 302:

[0038] Step 301: Input one of the feature maps into the spatial attention module and the channel attention module respectively.

[0039] Step 302: Combine the features corrected by the spatial attention module and the features corrected by the channel attention module to obtain a corrected feature vector.

[0040] Reference Figure 3 , the corrected feature vector is a high-dimensional feature.

[0041] In the process of feature extraction, the introduction of spatial attention module and channel attention module is the key to improving feature quality. The main function of the spatial attention module is to focus on the key information areas in the image, and emphasize those parts that play a key role in change monitoring by calculating the importance scores of each area. This enables the model to effectively identify potential changes in complex scenes. On the other hand, the channel attention module further enhances the model's ability to identify feature importance by dynamically adjusting the weights of each channel feature. By weighting the features of different channels, the model can more accurately extract the feature information that is most helpful for change monitoring. Such a feature correction mechanism effectively improves the model's sensitivity to subtle changes, so that change monitoring can still maintain high accuracy in the presence of noise and complex backgrounds. The optimization of this structure not only improves the richness and accuracy of feature expression, but also lays a good foundation for the final similarity analysis, making each step of change monitoring more reliable.

[0042] The above-mentioned DenseNet network, spatial attention module and channel attention module can constitute a multi-feature fusion change monitoring network based on DenseNet. The dual-image channel feature fusion change monitoring network based on DenseNet combines the advantages of DenseNet and aims to improve the accuracy and robustness of change monitoring. DenseNet is well-known for its efficient feature reuse mechanism. Through dense connection layers, the network can better transfer gradients during training and promote effective learning of features. In the change monitoring task, the model needs to deal with complex background changes and slight differences in target features. Therefore, fusing multiple features is the key. The network integrates features from different levels and makes full use of high-level semantic information and low-level detail information, so that the network's perception of changes can be improved. In addition, the DenseNet-based architecture enables the network to have better generalization ability and can adapt to the change monitoring needs in different scenarios. Compared with traditional change monitoring methods, this feature fusion strategy can more accurately capture the changes of target objects and provide more reliable monitoring results, thus showing obvious advantages in remote sensing image applications.

[0043] In another exemplary embodiment of the present application, the output high-dimensional features are subjected to similarity analysis to obtain the final change monitoring results. By analyzing through multiple similarity measurement methods, the change pattern can be automatically identified to ensure that efficient and accurate change monitoring performance can be maintained in a complex environment and provide reliable decision support.

[0044] The high-dimensional features after feature correction are subjected to similarity analysis to identify the changed areas between images. This process usually involves a variety of similarity measurement methods, such as cosine similarity, Euclidean distance, etc. By comparing the extracted feature vectors, the model can effectively monitor the differences between the two images. This step not only depends on the results of high-quality feature extraction and correction in the early stage, but also reflects the efficiency and accuracy of the entire process. Through similarity analysis, the model can automatically identify different change patterns, whether it is land use changes, new building construction, or the impact of natural disasters. This efficient change monitoring capability makes the present invention have a wide range of application potential in the fields of remote sensing image analysis, environmental monitoring, etc. Finally, after a series of processing and analysis, the obtained change monitoring results will provide an important basis for decision support, meet the actual needs of different application scenarios, and ensure that in complex and dynamic environments, efficient and accurate change monitoring performance can still be maintained.

[0045] Exemplarily, the objects to be monitored are human activities in terrestrial nature reserves, mainly including woodlands, grasslands, wetlands, houses, roads and mining land.

[0046] The advantages of this application method are as follows:

[0047] (1) Data level: Through the super-resolution network, low-resolution images are converted into higher-resolution images, thereby improving the clarity and detail of the images. The super-resolution network generates high-quality high-resolution images by learning a large number of high- and low-resolution image pairs using deep learning technology. This improves the spatial resolution of high-resolution images and the clarity and recognition efficiency of monitored objects.

[0048] (2) Methodology: By introducing a dual attention mechanism of spatial attention and channel attention, the model's ability to identify important features of the target can be significantly improved. Compared with traditional methods, the dual attention mechanism can effectively reduce feature interference and enhance the model's sensitivity to target features, thereby improving the performance of tasks such as image classification and target monitoring. This mechanism not only improves the performance of the model, but also provides a visual basis for understanding the model's decision-making process, making the final result more interpretable.

[0049] Based on the same inventive concept, the embodiment of the present application also provides a terrestrial nature reserve human activity remote sensing change monitoring device for implementing the above-mentioned terrestrial nature reserve human activity remote sensing change monitoring method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above-mentioned method, so the specific limitations in one or more embodiments of the terrestrial nature reserve human activity remote sensing change monitoring device provided below can refer to the above-mentioned limitations on the terrestrial nature reserve human activity remote sensing change monitoring method, and will not be repeated here.

[0050] In an exemplary embodiment, a device for remote sensing changes in human activities in a terrestrial nature reserve is provided, comprising: a resolution conversion module, a feature extraction module, a feature correction module and a similarity analysis module.

[0051] The resolution conversion module is used to input two remote sensing images of the object to be monitored at different time phases into a super-resolution network to obtain two super-resolution images at different time phases. The object to be monitored is human activities in a terrestrial nature reserve.

[0052] The feature extraction module is used to use the first DenseNet network to extract features from one super-resolution image, and use the second DenseNet network to extract features from another super-resolution image, so as to obtain two feature images in different phases.

[0053] A feature correction module is used to use the first dual attention module to correct the features of a feature map, and use the second dual attention module to correct the features of another feature map, so as to obtain feature vectors after correction in different time phases; the first dual attention module and the second dual attention module both adopt spatial attention mechanism and channel attention mechanism.

[0054] The similarity analysis module is used to perform similarity analysis on the corrected feature vectors at different time phases to obtain the change area between two remote sensing images of the monitored object at different time phases.

[0055] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 5 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the change area between two remote sensing images of the monitored object in different phases. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for remote sensing change monitoring of human activities in a terrestrial nature reserve is implemented.

[0056] Those skilled in the art will understand that Figure 5The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components. In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0057] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0058] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0059] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0060] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0061] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0062] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0063] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for monitoring changes in human activities in terrestrial nature reserves through remote sensing, characterized in that: The method for monitoring changes in human activities in terrestrial nature reserves through remote sensing includes: Inputting two remote sensing images of the object to be monitored at different time phases into a super-resolution network respectively to obtain two super-resolution images at different time phases; the object to be monitored is human activities in a terrestrial nature reserve; Use the first DenseNet network to extract features from one super-resolution image, and use the second DenseNet network to extract features from another super-resolution image, to obtain two feature images at different phases; A first dual attention module is used to perform feature correction on a feature map, and a second dual attention module is used to perform feature correction on another feature map to obtain feature vectors after correction in different time phases; both the first dual attention module and the second dual attention module adopt a spatial attention mechanism and a channel attention mechanism; The similarity analysis is performed on the corrected feature vectors at different time phases to obtain the change area between the two remote sensing images of the monitored object at different time phases.

2. The method for monitoring changes in human activities in terrestrial nature reserves by remote sensing according to claim 1, characterized in that: The first DenseNet network and the second DenseNet network adopt a shared weight mechanism.

3. The method for monitoring changes in human activities in terrestrial nature reserves by remote sensing according to claim 1, characterized in that: The first dual attention module and the second dual attention module both include: a spatial attention module and a channel attention module.

4. The method for monitoring changes in human activities in terrestrial nature reserves by remote sensing according to claim 3, characterized in that: The first dual attention module is used to perform feature correction on a feature map, specifically including: Inputting one of the feature maps into a spatial attention module and a channel attention module respectively; The features corrected by the spatial attention module and the features corrected by the channel attention module are combined to obtain a corrected feature vector.

5. The method for monitoring changes in human activities in terrestrial nature reserves by remote sensing according to claim 1, characterized in that: The similarity measurement method used in the similarity analysis includes at least cosine similarity or Euclidean distance.

6. The method for monitoring changes in human activities in terrestrial nature reserves by remote sensing according to claim 1, characterized in that: The objects to be monitored also include: woodlands, grasslands, wetlands, houses, roads and mining land.

7. A remote sensing change monitoring device for human activities in terrestrial nature reserves, characterized in that: The remote sensing change monitoring device for human activities in terrestrial nature reserves comprises: A resolution conversion module, for inputting two remote sensing images of an object to be monitored at different time phases into a super-resolution network respectively, to obtain two super-resolution images at different time phases; the object to be monitored is human activities in a terrestrial nature reserve; A feature extraction module is used to extract features from one super-resolution image using a first DenseNet network, and to extract features from another super-resolution image using a second DenseNet network, thereby obtaining two feature images at different phases; A feature correction module, used to perform feature correction on a feature map using a first dual attention module, and to perform feature correction on another feature map using a second dual attention module, to obtain feature vectors after correction in different time phases; both the first dual attention module and the second dual attention module adopt a spatial attention mechanism and a channel attention mechanism; The similarity analysis module is used to perform similarity analysis on the corrected feature vectors at different time phases to obtain the change area between two remote sensing images of the monitored object at different time phases.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for remote sensing change monitoring of human activities in terrestrial nature reserves as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for monitoring changes in human activities in terrestrial nature reserves through remote sensing is implemented as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for monitoring changes in human activities in terrestrial nature reserves through remote sensing is implemented as described in any one of claims 1 to 6.

Images