Landslide detection method and device, program product and electronic equipment

Abstract

The invention provides a landslide detection method and device, a program product and electronic equipment, and relates to the technical field of computers. The method comprises the following steps: determining a to-be-detected remote sensing image, and then performing feature extraction on the to-be-detected remote sensing image through a backbone network of a landslide detection model to obtain a plurality of initial feature maps with different sizes; and through the neck network of the landslide detection model, carrying out fusion processing on the plurality of initial feature maps with different sizes to obtain a plurality of fusion feature maps with different sizes, and then through the receptive field attention sub-module of the landslide detection model, carrying out feature processing on the plurality of fusion feature maps with different sizes to obtain target receptive field features. Furthermore, the target receptive field features are detected and processed through the detection network of the landslide detection model to obtain a landslide detection frame and a landslide category label, so that the accuracy of detecting the landslide in the remote sensing image is improved to a certain extent.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to a landslide detection method, device, program product, and electronic device. Background Art

[0002] Landslides are natural geological disasters caused by a variety of factors such as earthquakes, rainfall, and human activities. They seriously threaten human life and property safety. Therefore, how to quickly and accurately detect landslides to provide great convenience and assistance for disaster relief and management has become an urgent problem to be solved.

[0003] In optical images, due to the destruction of the integrity of the original terrain, landslides usually exhibit the following characteristics: (1) exposed soil partially covered by vegetation; (2) broken surface; (3) rough and uneven texture; (4) discontinuous landscape pattern. Moreover, over time, due to factors such as vegetation regeneration and surface weathering, the aforementioned distinct surface features of landslides gradually weaken. For such landslides, existing detection methods mainly extract landslide features from optical images and rely on the expertise and experience of landslide experts to manually analyze remote sensing images to delineate the location and boundaries of the landslide to determine the landslide detection results. Summary of the Invention

[0004] However, the aforementioned method of determining landslide locations and boundaries based on manual analysis is not only time-consuming and labor-intensive, but the landslide detection results determined based on this method are also affected by the subjective judgment of individual experts, resulting in low accuracy in landslide judgment.

[0005] The present disclosure provides a landslide detection method, a landslide detection device, a computer program product, and an electronic device to improve the accuracy of landslide detection.

[0006] According to a first aspect of the present disclosure, a landslide detection method is provided, the method comprising: Determine the remote sensing image to be detected; Performing feature extraction on the remote sensing image to be detected through a backbone network of the landslide detection model to obtain a plurality of initial feature maps of different sizes; and performing fusion processing on the plurality of initial feature maps of different sizes through a neck network of the landslide detection model to obtain a plurality of fused feature maps of different sizes; Performing feature processing on the multiple fused feature maps of different sizes through the receptive field attention submodule of the landslide detection model to obtain target receptive field features; The target receptive field features are detected and processed by the detection network of the landslide detection model to obtain a landslide detection frame and a landslide category label.

[0007] In a possible implementation, the receptive field attention submodule of the landslide detection model performs feature processing on the multiple fused feature maps of different sizes to obtain target receptive field features, including: Inputting the multiple fused feature maps of different sizes into the first sub-branch of the receptive field attention sub-module to obtain a first feature matrix; Inputting the multiple fused feature maps of different sizes into the second sub-branch of the receptive field attention sub-module to obtain a second feature matrix; Input the second feature matrix into the third sub-branch of the receptive field attention sub-module to obtain a third feature matrix; Determine target receptive field features based on the first feature matrix and the third feature matrix.

[0008] In a possible implementation, inputting the multiple fused feature maps of different sizes into the first sub-branch of the receptive field attention sub-module to obtain a first feature matrix includes: The multiple fused feature maps of different sizes are input into the nonlinear activation function of the receptive field attention sub-module to obtain the channel attention vector of the original feature corresponding to the fused feature map, and the channel attention vector is used as the first feature matrix.

[0009] In a possible implementation, inputting the multiple fused feature maps of different sizes into the second sub-branch of the receptive field attention sub-module to obtain a second feature matrix includes: Performing global average pooling and one-dimensional convolution processing on the multiple fused feature maps of different sizes to obtain a first sub-feature matrix; The first sub-feature matrix is processed using convolution kernels of different sizes and dilated convolution to obtain a second feature matrix.

[0010] In a possible implementation, determining target receptive field features according to the first feature matrix and the third feature matrix includes:

[0011] in, X out Used to characterize the target receptive field characteristics; among them, , X F Used to characterize the second characteristic matrix, X Used to characterize the feature matrix corresponding to the fused feature map; X a Used to characterize the third characteristic matrix; Used to characterize the first characteristic matrix; Re LU ( X a) is used to represent the application of the ReLU activation function to each element in the third feature matrix.

[0012] In a possible implementation, the landslide detection model is trained based on the following method: Collecting a remote sensing image data set and dividing the remote sensing image data set into a training set, a validation set, and a test set; Training a preset landslide detection model according to the training set to obtain multiple prediction information; Determining comparison information between the plurality of prediction information and the labeled information in the training set through a loss function, and adjusting the preset landslide detection model according to the comparison information to obtain an adjusted preset landslide detection model; Adjusting the hyperparameters in the adjusted preset landslide detection model according to the validation set, and obtaining an updated preset landslide detection model when it is determined that the validation set loss reaches a first preset condition; The updated preset landslide detection model is evaluated according to the test set, and when it is determined that the evaluation result meets the second preset condition, the landslide detection model is obtained.

[0013] In one possible implementation, the loss function is determined based on the following formula:

[0014] in, N 1 is used to represent the prediction information, N 2 is used to represent the annotation information, C 2 is used to represent a constant related to the training set, Used to characterize the second-order Wasserstein distance between the predicted information and the labeled information; NWD ( N 1 , N 2 ) is used to characterize the similarity between the predicted information and the labeled information.

[0015] According to a second aspect of the present disclosure, a landslide detection device is provided, the device comprising: A determination unit, used for determining a remote sensing image to be detected; The detection unit is used to extract features from the remote sensing image to be detected through the backbone network of the landslide detection model to obtain multiple initial feature maps of different sizes; and to fuse the multiple initial feature maps of different sizes through the neck network of the landslide detection model to obtain multiple fused feature maps of different sizes; to perform feature processing on the multiple fused feature maps of different sizes through the receptive field attention submodule of the landslide detection model to obtain target receptive field features; and to detect and process the target receptive field features through the detection network of the landslide detection model to obtain a landslide detection frame and a landslide category label.

[0016] In a possible implementation, the detection unit is specifically configured to: Inputting the multiple fused feature maps of different sizes into the first sub-branch of the receptive field attention sub-module to obtain a first feature matrix; Inputting the multiple fused feature maps of different sizes into the second sub-branch of the receptive field attention sub-module to obtain a second feature matrix; Input the second feature matrix into the third sub-branch of the receptive field attention sub-module to obtain a third feature matrix; Determine target receptive field features based on the first feature matrix and the third feature matrix.

[0017] In a possible implementation, the detection unit is specifically configured to: The multiple fused feature maps of different sizes are input into the nonlinear activation function of the receptive field attention sub-module to obtain the channel attention vector of the original feature corresponding to the fused feature map, and the channel attention vector is used as the first feature matrix.

[0018] In a possible implementation, the detection unit is specifically configured to: Performing global average pooling and one-dimensional convolution processing on the multiple fused feature maps of different sizes to obtain a first sub-feature matrix; The first sub-feature matrix is processed using convolution kernels of different sizes and dilated convolution to obtain a second feature matrix.

[0019] In a possible implementation, the detection unit is specifically configured to:

[0020] in, X out Used to characterize the target receptive field characteristics; among them, , X F Used to characterize the second characteristic matrix, X A feature matrix for representing the fused feature map; Xa Used to characterize the third characteristic matrix; Used to characterize the first characteristic matrix; Re LU ( X a ) is used to represent the application of the ReLU activation function to each element in the third feature matrix.

[0021] In a possible implementation, the device further includes a training unit configured to: Collecting a remote sensing image data set and dividing the remote sensing image data set into a training set, a validation set, and a test set; Training a preset landslide detection model according to the training set to obtain multiple prediction information; Determining comparison information between the plurality of prediction information and the labeled information in the training set through a loss function, and adjusting the preset landslide detection model according to the comparison information to obtain an adjusted preset landslide detection model; Adjusting the hyperparameters in the adjusted preset landslide detection model according to the validation set, and obtaining an updated preset landslide detection model when it is determined that the validation set loss reaches a first preset condition; The updated preset landslide detection model is evaluated according to the test set, and when it is determined that the evaluation result meets the second preset condition, the landslide detection model is obtained.

[0022] In one possible implementation, the loss function is determined based on the following formula:

[0023] in, N 1 is used to represent the prediction information, N 2 is used to represent the annotation information, C 2 is used to represent a constant related to the training set, Used to characterize the second-order Wasserstein distance between the predicted information and the labeled information; NWD ( N 1 , N 2 ) is used to characterize the similarity between the predicted information and the labeled information.

[0024] According to a third aspect of the present disclosure, a computer program product is provided, comprising a computer program, wherein when the computer program is executed by a processor, the method of the first aspect and possible implementations thereof are implemented.

[0025] According to a fourth aspect of the present disclosure, an electronic device is provided, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute the method of the above-mentioned first aspect and its possible implementation methods by executing the executable instructions.

[0026] The technical solution disclosed in this disclosure has the following beneficial effects: In an embodiment of the present disclosure, a remote sensing image to be detected can be determined, and then feature extraction can be performed on the remote sensing image to be detected through the backbone network of the landslide detection model to obtain multiple initial feature maps of different sizes; and, the multiple initial feature maps of different sizes can be fused through the neck network of the landslide detection model to obtain multiple fused feature maps of different sizes; the receptive field attention submodule of the landslide detection model can be used to perform feature processing on the multiple fused feature maps of different sizes to obtain target receptive field features; and the target receptive field features can be detected and processed through the detection network of the landslide detection model to obtain a landslide detection frame and a landslide category label. It can be seen that the present disclosure adopts a landslide detection model, which can not only automatically detect the remote sensing image to be detected, avoiding errors caused by manual analysis and improving the accuracy of detection, but also combines the receptive field attention submodule to perform landslide detection on the remote sensing image to be detected based on the features determined by deep mining of the remote sensing image to be detected, further improving the accuracy of detection of the remote sensing image to be detected.

[0027] Other features and advantages of the present disclosure will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present disclosure. The purposes and other advantages of the present disclosure can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the following briefly introduces the drawings required for use in the embodiments of the present disclosure. Obviously, the drawings introduced below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0029] Figure 1 A schematic diagram of an application scenario in this exemplary embodiment is shown; Figure 2 A schematic diagram showing a process of obtaining a landslide detection model in this exemplary embodiment is shown; Figure 3 A schematic diagram showing a remote sensing image in an initial data set according to this exemplary embodiment; Figure 4 A schematic diagram showing a remote sensing image in a processed data set according to this exemplary embodiment; Figure 5 A schematic flow chart showing a landslide detection method in this exemplary embodiment is shown; Figure 6 A schematic diagram showing the structure of an efficient channel attention model in this exemplary embodiment is shown; Figure 7 A schematic diagram showing a landslide detection model in this exemplary embodiment; Figure 8 A schematic structural diagram of a landslide detection device according to this exemplary embodiment is shown; Figure 9 A schematic structural diagram of an electronic device in this exemplary embodiment is shown. DETAILED DESCRIPTION

[0030] In order to make the purpose, technical solutions and advantages of the present disclosure clearer, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present disclosure. In the absence of conflict, the embodiments in the present disclosure and the features in the embodiments can be arbitrarily combined with each other. In addition, although the logical order is shown in the flow chart, in some cases, the steps shown or described can be performed in an order different from that here.

[0031] The terms "comprises" and "comprising" and any variations thereof in the specification and claims of this disclosure are intended to cover non-exclusive protection. For example, a process, method, system, product, or apparatus comprising a series of steps or elements is not limited to the listed steps or elements, but may optionally include steps or elements not listed, or may optionally include other steps or elements inherent to the process, method, product, or apparatus.

[0032] In the embodiments of the present disclosure, one or more, "many" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: the existence of A alone, the existence of A and B at the same time, and the existence of B alone, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, c can be single or plural.

[0033] It should be noted that the terms "first", "second", "third", etc. in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order, sequence, size and priority. For example, the first sub-branch, the second sub-branch and the third sub-branch in the embodiment of the present disclosure are only used to distinguish different branches. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0034] The following describes exemplary embodiments of the present disclosure in conjunction with the accompanying drawings. The accompanying drawings are schematic diagrams of the present disclosure and are not necessarily drawn to scale. Some of the block diagrams shown in the accompanying drawings may be functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, or in hardware modules or integrated circuits, or in networks, processors or microcontrollers. The embodiments can be implemented in various forms and should not be construed as being limited to the examples set forth herein. The features, structures or characteristics described in the present disclosure can be combined in one or more embodiments in any suitable manner. In the description below, many specific details are provided to provide a full description of the embodiments of the present disclosure. However, those skilled in the art will appreciate that one or more specific details may be omitted when implementing the technical solutions of the present disclosure, or that other methods, components, devices, steps, etc. may be used to replace one or more specific details.

[0035] It should be noted that in the embodiments of this disclosure, certain software, components, models, and other existing industry solutions may be mentioned. These should be considered exemplary and are intended solely to illustrate the feasibility of implementing the technical solutions of this disclosure. This does not mean that the applicant has or will necessarily use such solutions. The collection, dissemination, and use of data in this technical solution comply with relevant national laws and regulations.

[0036] In recent years, deep learning methods have been increasingly applied to remote sensing, including the automated detection of landslides. This process typically involves two steps: first, landslide areas are labeled based on prior knowledge, and then the labeled data is used to train a deep learning model to detect landslides in other areas.

[0037] Specifically, to improve detection performance, a contextual enhancement module consisting of unrolled convolutions is introduced into the regression branch of the decoupled head. This improved decoupled head replaces the coupled head in YOLOv5s to enhance model effectiveness. Furthermore, a multi-scale attention model based on the Swin Transformer (a visual model based on the Transformer architecture that excels in computer vision tasks such as image classification and object detection, with a core feature of improving computational efficiency and capturing local features through a shifted window mechanism) is used to comprehensively capture and learn multi-scale landslide features. Furthermore, the Multi-Scale Feature Fusion Lightweight Neural Network (MFFLnet) employs a deep transfer learning (DTL) strategy, enabling MFFLnet to leverage prior landslide knowledge from the source domain to mitigate overfitting.

[0038] However, these aforementioned models show inconsistent performance when dealing with small objects, objects of extreme shapes (e.g., slender, narrow, or tall), and datasets containing different object types in the same class.

[0039] Therefore, the current research also proposed corresponding methods to address specific landslide challenges. First, combining morphological edge recognition with the Swin Transformer deep learning model enhances boundary delineation, effectively addressing the problems of irregular boundaries and discretized features in landslide detection.

[0040] However, due to interfering factors such as vegetation cover, models often struggle to fully understand the optical properties of landslides. To address this issue, the texture features of high-resolution imagery can be leveraged, combined with auxiliary features such as the Normalized Difference Vegetation Index (NDVI) and the Gray-Level Co-Occurrence Matrix (GLCM), to improve model performance.

[0041] However, these methods are only targeted at specific areas, and their performance in large-scale landslide detection has not been fully verified, resulting in significant limitations.

[0042] To explore more effective landslide detection solutions, a network that simultaneously focuses on local and global features of landslides is developed by combining the advantages of convolutional neural networks (CNNs) and transformer models. Despite their advantages, transformer-based models and their improved variants typically require large datasets and long training times to achieve optimal performance.

[0043] In view of this, an exemplary embodiment of the present disclosure provides a landslide detection method, by which a remote sensing image to be detected can be determined, and then feature extraction is performed on the remote sensing image to be detected through the backbone network of the landslide detection model to obtain multiple initial feature maps of different sizes; and the multiple initial feature maps of different sizes are fused and processed through the neck network of the landslide detection model to obtain multiple fused feature maps of different sizes; the multiple fused feature maps of different sizes are feature processed through the receptive field attention submodule of the landslide detection model to obtain target receptive field features; and the target receptive field features are detected and processed through the detection network of the landslide detection model to obtain a landslide detection frame and a landslide category label. It can be seen that the present disclosure adopts a landslide detection model, which can not only automatically detect the remote sensing image to be detected, avoiding errors caused by manual analysis and improving detection accuracy, but also combines the receptive field attention submodule to perform landslide detection on the remote sensing image to be detected based on the features determined by deep mining of the remote sensing image to be detected, further improving the accuracy of detection of the remote sensing image to be detected.

[0044] To better understand the technical solutions provided by the embodiments of the present disclosure, the following briefly introduces the application scenarios to which the technical solutions provided by the embodiments of the present disclosure are applicable. It should be noted that the application scenarios introduced below are only used to illustrate the embodiments of the present disclosure and are not limiting. In specific implementations, the technical solutions provided by the embodiments of the present disclosure can be flexibly applied according to actual needs.

[0045] See also Figure 1 As shown, Figure 1 This is an application scenario to which the technical solutions of the embodiments of the present disclosure can be applied. In this scenario diagram, a terminal device 110 and an electronic device 120 are included. The terminal device 110 and the electronic device 120 can be directly or indirectly connected to each other via one or more networks 130. Optionally, the application scenario may also include other electronic devices, which can, for example, perform linked processing based on the output results of the electronic device 120, but this is not limited in the embodiments of the present disclosure.

[0046] In an embodiment of the present disclosure, a user can send a remote sensing image to be detected based on the terminal device 110, and then the electronic device 120 can determine the remote sensing image to be detected; feature extraction is performed on the remote sensing image to be detected through the backbone network of the landslide detection model to obtain multiple initial feature maps of different sizes; and, the neck network of the landslide detection model is used to fuse the multiple initial feature maps of different sizes to obtain multiple fused feature maps of different sizes; the receptive field attention submodule of the landslide detection model is used to perform feature processing on the multiple fused feature maps of different sizes to obtain target receptive field features; the target receptive field features are detected and processed through the detection network of the landslide detection model to obtain a landslide detection frame and a landslide category label.

[0047] In the embodiments of the present disclosure, Figure 1 The terminal device 110 may be a mobile phone, a tablet computer (PAD), a personal computer (PC), a smart TV, a smart watch, a smart speaker, a smart car device, a wearable device, etc., but is not limited thereto.

[0048] In the embodiment of the present disclosure, the electronic device 120 can be a server. The server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers. It can also be a cloud server or cloud server cluster that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms, but is not limited to these.

[0049] Of course, the method provided in the embodiment of the present disclosure is not limited to Figure 1 The application scenario shown can also be used in other possible application scenarios, such as an application scenario where the landslide detection method is implemented only by the electronic device 120, and the embodiments of the present disclosure are not limited thereto.

[0050] To further illustrate the technical solutions provided by the embodiments of the present disclosure, this is described in detail below in conjunction with the accompanying drawings and specific implementation methods. Although the embodiments of the present disclosure provide the method operation steps as shown in the following embodiments or drawings, more or fewer operation steps may be included in the method based on routine or no creative labor. In steps where there is no necessary causal relationship logically, the execution order of these steps is not limited to the execution order provided by the embodiments of the present disclosure. In the actual processing process or when the device is executed, the method may be executed in the order of the methods shown in the embodiments or drawings or in parallel.

[0051] The following combination Figure 5The method flow chart shown in FIG. 1 illustrates the landslide detection method in the embodiment of the present disclosure. Figure 5 The steps shown can be performed as follows Figure 1 The electronic device 120 is shown performing.

[0052] In the embodiment of the present disclosure, it is possible to determine the collection of a remote sensing image dataset, divide the remote sensing image dataset into a training set, a validation set, and a test set, and then use the training set, the validation set, and the test set to train a preset landslide detection model to obtain a landslide detection model. The landslide detection model is then used to detect new remote sensing images to be detected. Before introducing the landslide detection method, the following first introduces the process of obtaining the landslide detection model in the embodiment of the present disclosure. Please refer to Figure 2 , Figure 2 Schematic diagram of the process of obtaining a landslide detection model in an embodiment of the present disclosure.

[0053] Step 201: Collect a remote sensing image dataset and divide the remote sensing image dataset into a training set, a validation set, and a test set.

[0054] In the disclosed embodiments, remote sensing image datasets can be collected from different regions. For example, remote sensing image datasets from three regions can be collected: the Bijie landslide dataset, the Luding landslide dataset, and the Southwest landslide dataset. The first three datasets are derived from different satellite imagery and cover landslide events in different regions. In other words, the remote sensing image datasets in the disclosed embodiments have a relatively rich foundation of data, providing a good foundation for subsequent identification of landslide events in different regions.

[0055] See also Figure 3 As shown, Figure 3 A schematic diagram of a remote sensing image of a Bijie landslide dataset provided by an embodiment of the present disclosure. Specifically, Figure 3 The Bijie landslide dataset shown here is from the Bijie Landslide Dataset published by Wuhan University. The dataset covers the entire city of Bijie, Guizhou Province, China, with a total area of 26,853 square kilometers. The area has an average elevation of 1,600 meters and is surrounded by mountains and rivers. Due to its steep slopes, fragile ecosystem, and unstable geology, the region is one of the most landslide-prone areas in China. The dataset contains 770 optical remote sensing images of landslide samples, acquired by the TripleSat satellite between May and August 2018. The images are composed of three primary RGB color bands and have a resolution of 0.8 meters.

[0056] In the embodiment of the present disclosure, after obtaining a remote sensing data image set, considering that the images in the remote sensing data image set have different sizes, the data in the remote sensing data image set can be standardized to ensure that the mean of the input data is 0 and the standard deviation is 1. Optionally, a letterbox filling strategy (LetterBox strategy) can be used to uniformly adjust all samples to 320×320 pixels. The LetterBox strategy is an image processing technology used to proportionally adjust images of different sizes to a fixed size while maintaining the original aspect ratio of the image and avoiding geometric distortion of the image. For example, see Figure 4 As shown, Figure 4 , which shows the remote sensing images in the remote sensing data image set after adjustment according to the LetterBox strategy.

[0057] Optionally, the remote sensing data image set can be randomly divided into three groups according to a preset ratio, and used for training, validation, and testing, respectively, to obtain a training set, a validation set, and a test set. The preset ratio is, for example, 8:1:1, but other ratios are also possible and are not limited in the present embodiment.

[0058] Step 202: Train a preset landslide detection model according to the training set to obtain multiple prediction information.

[0059] In the disclosed embodiments, the preset landslide detection model can be an improved YOLOv11 model that incorporates receptive field attention, normalized Wasserstein distance, and the removal of the CrossStage Partial with Pyramid Squeeze Attention (C2PSA) module. This means that while retaining the advantages of the YOLOv11 model, the preset landslide detection model provided in the disclosed embodiments reconstructs the network structure to enhance feature extraction and small object recognition capabilities.

[0060] Step 203: determining comparison information between the plurality of prediction information and the labeled information in the training set through a loss function, and adjusting the preset landslide detection model according to the comparison information to obtain an adjusted preset landslide detection model.

[0061] In one possible implementation, the loss function is determined based on the following formula 1: Formula 1 in, N 1 is used to represent the prediction information, N 2 is used to represent the annotation information, C 2 is used to represent a constant related to the training set, Used to characterize the second-order Wasserstein distance between the predicted information and the labeled information; NWD ( N 1 , N 2 ) is used to characterize the similarity between the predicted information and the labeled information. Optionally, Usually set to the mean absolute value of the training set.

[0062] In the disclosed embodiment, NWD first models the bounding box as a two-dimensional Gaussian distribution, and then uses the normalized Wasserstein distance to measure the similarity of the derived Gaussian distributions.

[0063] Specifically, given two two-dimensional Gaussian distributions and , the second-order Wasserstein distance between them is defined as: Formula 2.

[0064] To convert this distance into a similarity metric between 0 and 1, NWD uses an exponentially normalized loss function. In the disclosed embodiments, this loss function design makes NWD insensitive to targets of varying scales, making it more suitable for measuring the similarity between the true and predicted values of small targets in remote sensing imagery within the model. In other words, the preset landslide detection model in this disclosure utilizes a normalized Wasserstein distance loss instead of the traditional intersection-over-union loss, improving the preset landslide detection model's ability to distinguish between positive and negative samples and promoting more efficient convergence of the preset landslide detection model.

[0065] Step 204: adjusting the hyperparameters in the adjusted preset landslide detection model according to the validation set, and obtaining an updated preset landslide detection model when it is determined that the validation set loss reaches a first preset condition.

[0066] Step 205: Evaluate the updated preset landslide detection model according to the test set, and when it is determined that the evaluation result meets the second preset condition, obtain the landslide detection model.

[0067] In the embodiment of the present disclosure, the first preset condition and the second preset condition can be set accordingly based on actual implementation, and this is not limited in the embodiment of the present disclosure.

[0068] Furthermore, in the embodiment of the present disclosure, after obtaining the landslide detection model, the remote sensing image to be detected can be detected according to the landslide detection model. For details, please refer to Figure 5 The flowchart shown.

[0069] Step 501: Determine the remote sensing image to be detected.

[0070] In the embodiments of the present disclosure, a user may send the remote sensing image to be detected to the electronic device through a terminal device, so that the electronic device determines the remote sensing image to be detected. Alternatively, the electronic device may determine the remote sensing image to be detected from a set of collected remote sensing images based on preset rules. This is not limited in the embodiments of the present disclosure.

[0071] Step 502: extract features from the remote sensing image to be detected through the backbone network of the landslide detection model to obtain multiple initial feature maps of different sizes, and fuse the multiple initial feature maps of different sizes through the neck network of the landslide detection model to obtain multiple fused feature maps of different sizes.

[0072] Step 503: Perform feature processing on multiple fused feature maps of different sizes through the receptive field attention submodule of the landslide detection model to obtain target receptive field features.

[0073] In an embodiment of the present disclosure, multiple fused feature maps of different sizes can be input into the first sub-branch of the Receptive-Field Attention (RFA) module to obtain a first feature matrix; the first feature matrix can be input into the second sub-branch of the receptive field attention sub-module to obtain a second feature matrix; the second feature matrix can be input into the third sub-branch of the receptive field attention sub-module to obtain a third feature matrix; and the target receptive field features can be determined based on the first feature matrix and the third feature matrix.

[0074] Optionally, inputting multiple fused feature maps of different sizes into a first sub-branch of a receptive field attention sub-module to obtain a first feature matrix includes: inputting the fused feature map into a nonlinear activation function of the receptive field attention sub-module, obtaining channel attention vectors of the original features corresponding to the multiple fused feature maps of different sizes, and using the channel attention vectors as the first feature matrix. For example, the first sub-branch processes the fused feature map using a Sigmoid function to obtain the channel attention vectors of the original features corresponding to the fused feature map, and generates corresponding weights to obtain the first feature matrix.

[0075] Optionally, multiple fused feature maps of different sizes are input into the second sub-branch of the receptive field attention sub-module to obtain a second feature matrix, including: performing global average pooling and one-dimensional convolution on the multiple fused feature maps of different sizes to obtain a first sub-feature matrix; and processing the first sub-feature matrix using convolution kernels of different sizes and dilated convolution to obtain a second feature matrix.

[0076] Specifically, the second sub-branch consists of Efficient Channel Attention (ECA) and Receptive Field Block (RFB) to fully extract feature information. ECA consists of global average pooling and one-dimensional convolution modules. Specifically, for input X, ECA is calculated as shown in Formula 3 and Formula 4 below: Formula 3 Formula 4 in, F ECA ( X ) is used to represent efficient channel attention processing on the input X, σ is used to represent the Sigmoid function, GAP(X) Often indicates input X Perform a global average pooling operation, Used to represent the channel attention vector, X Used to represent multiple fusion feature maps of different sizes, Y Used to characterize the first sub-feature matrix, Conv1D Used to represent a one-dimensional convolution, the size of the convolution kernel is k, k By channel dimension C Adaptive determination, where ，k The calculation formula is shown in Formula 5 below: Formula 5 Among them, |∙|odd means the nearest odd number to |∙|, γ and b is a hyperparameter. Optionally, in the embodiment of the present disclosure, γ and b The values of are set to 2 and 1 respectively.

[0077] For example, see Figure 6 As shown, k When it is equal to 5, ECA can perform Figure 6 The processing shown in FIG5 is to collect, for the input X with a height of H, a width of W, and a channel dimension of C, first perform a global average pooling operation, then perform a one-dimensional convolution operation with a convolution kernel of 5, and then process the result of the convolution operation through the Sigmoid function to obtain the first sub-feature matrix.

[0078] In the disclosed embodiment, RFB adopts a multi-branch pool, in which different branches use convolution kernels of different sizes to respond to receptive fields of different sizes. In other words, RFB designs different padding and expansion to process small, medium and large range feature information respectively. Finally, the three branches of RFB are spliced together to form a complete receptive field feature as the second feature matrix. In addition, the relationship between the receptive field size and the eccentricity is also considered in RFB, which is conducive to highlighting the importance of the area close to the center, improving the insensitivity to small spatial displacements, and helping the network focus on identifying high-level abstract features without being affected by specific spatial positions.

[0079] That is, in RFA, the feature map is first processed by ECA, and then, RFB reshapes the feature map sufficiently to highlight the importance of the region close to the center, thereby allowing the landslide detection model to focus on identifying high-level abstract features.

[0080] In the embodiment of the present disclosure, after obtaining the first feature matrix, the second feature matrix, and the third feature matrix, the target receptive field feature can be determined based on the following formula 6: Formula 6 in, X out Used to characterize the target receptive field characteristics; among them, , X F Used to characterize the second characteristic matrix, X A feature matrix for representing the fused feature map; X a Used to characterize the third characteristic matrix; Used to characterize the first characteristic matrix; Re LU ( X a ) is used to represent the application of the Rectified Linear Unit (ReLU) activation function to each element in the third feature matrix.

[0081] As can be seen, receptive field attention (RFA) is introduced at three locations on the neck of the landslide detection model and connected to the detection head to detect small, medium, and large objects. In this way, RFA simulates the human visual receptive field by combining multi-scale convolution operations and channel attention, enabling the landslide detection model to better capture multi-scale landslide features and improve the detection accuracy of the remote sensing imagery to be tested.

[0082] Step 604: Detect and process the target receptive field features through the detection network of the landslide detection model to obtain a landslide detection frame and a landslide category label.

[0083] In the disclosed embodiment, the detection network of the landslide detection model is responsible for predicting the location, category, and confidence of each bounding box. Specifically, the detection network simultaneously predicts the category of each bounding box, uses a softmax function to map the feature map to the category probability distribution, removes redundant bounding boxes through non-maximum suppression (NMS), and retains the bounding boxes with the highest confidence. When training the detection network, the normalized Wasserstein distance (NWD) loss function can be used instead of the traditional intersection over union (IoU) loss function, thereby improving the detection performance of small objects.

[0084] In the disclosed embodiments, by simulating the human visual receptive field and combining multi-scale convolution operations with a channel attention mechanism, complex and multi-scale landslide features can be effectively extracted. This allows the landslide detection model to excel in processing landslide targets of varying sizes and shapes, particularly with significantly improved detection capabilities for small targets. Furthermore, the disclosed embodiments introduce NWD into the loss function, replacing the traditional intersection-over-union (IoU) calculation method. As a result, the landslide detection model provided by the disclosed embodiments performs well when processing small targets, more accurately calculating the similarity between the predicted and true frames, reducing errors caused by small targets and improving the model's detection accuracy for small targets.

[0085] In order to better introduce the solution provided by the embodiment of the present disclosure, see Figure 7 As shown, the landslide detection scheme is introduced below with a specific example.

[0086] See also Figure 7It can be seen that the landslide detection model provided by the embodiment of the present disclosure uses CBS (an abbreviation of a basic module, consisting of three core components: C: Convolution (convolution layer); B: Batch Normalization (batch normalization); S: SiLU / Swish (activation function, usually SiLU, i.e. Sigmoid Linear Unit)) for basic feature extraction, C3k2 / C3k (where C3 (also known as CSP Bottleneck) is a variant of Cross Stage Partial Network (CSPNet), mainly composed of the following three parts: trunk branch: containing multiple residual blocks (Bottleneck); shortcut branch: directly connecting the input to retain the original features; final fusion: merging the features of the two branches through the Concat operation. C3k or C3k2 is a variant of C3) to construct a multi-branch residual block to enrich feature expression, and realize cross-scale feature interaction through "upsampling + feature fusion"; uses the fast spatial pyramid pooling module (Spatial Pyramid Pooling – Fast, SPPF) to compress features, RFA to enhance attention, and finally uses NWD loss to optimize the detection task. Furthermore, the C2PSA block has been removed from YOLOv11, simplifying the model structure and improving overall efficiency. These improvements enable the landslide detection model provided by the disclosed embodiments to achieve excellent landslide detection performance in a variety of scenarios. These scenarios, such as multi-scale targets, complex optical features, and limited sample conditions, are not limited in the disclosed embodiments.

[0087] In the embodiment of the present disclosure, the remote sensing image to be detected can be input into the backbone network in the landslide detection model, and then the backbone network performs downsampling (i.e., reducing the size) and feature extraction processing on the remote sensing image to be detected to generate feature maps of different levels, such as Figure 7 The initial feature maps are obtained by multiplying the scales (320×, 160×, etc.) shown in the figure. Based on the neck network in the landslide detection model, the multi-scale feature maps output by the backbone network are fused across layers to obtain three fused feature maps of different sizes.

[0088] In the embodiment of the present disclosure, three fused feature maps of different sizes can be input into the first sub-branch of the receptive field attention sub-module of the landslide detection model to obtain the first feature matrix, that is, Figure 7 The feature matrix corresponding to the leftmost branch in RFA. In addition, the fusion feature maps of three different sizes can be input into the second sub-branch of the receptive field attention sub-module (i.e. Figure 7The ECA in the middle branch of the RFA in the ensemble is used to obtain the first sub-feature matrix, and the first sub-feature matrices corresponding to the three fused feature maps of different sizes are input into different branch pools, and different branches use convolution kernels of different sizes to respond to receptive fields of different sizes.

[0089] For example, Figure 7 The convolutions in the left branch of the RFB shown in the figure are configured as 3×3 convolutions (dilation rate = 1, padding = 1). The left branch has the smallest receptive field and is suitable for detecting small objects and details (such as edges and textures). The convolutions in the middle branch of the RFB are configured as 3×3 convolutions (dilation rate = 2, padding = 2) followed by 3×3 convolutions (dilation rate = 1, padding = 1). The middle branch has a medium receptive field and is suitable for detecting medium-sized objects and semi-global semantics. Furthermore, the convolutions in the right branch of the RFB are configured as three layers of 3×3 convolutions with dilation rates of 1 → 2 → 3, and increasing padding. The left branch has the largest receptive field and is suitable for detecting large objects and global structures (such as object contours). Then, channel dimensionality reduction and feature fusion are performed on the left, middle, and right branches in the RFB to obtain the complete receptive field features, namely the second feature matrix.

[0090] In the disclosed embodiments, feature maps from different receptive fields can be directly added together in the spatial dimension, allowing the features of the second and first feature matrices to complement each other to determine a third feature matrix. The target receptive field features can then be determined based on the third and first feature matrices. Furthermore, the target receptive field features can be input into a detection network to obtain a landslide detection box and landslide category label.

[0091] During the specific implementation process, the inventors used high-resolution remote sensing image data in the experiment to verify the detection effect of the landslide detection method based on the landslide detection model provided by the embodiment of the present disclosure and other landslide detection methods. The verification results are shown in Table 1, for example. Obviously, the improved YOLOv11, i.e., the landslide detection model provided by the embodiment of the present disclosure, has the best comprehensive performance.

[0092] Table 1 Verification table of landslide detection model detection effect

[0093] Specifically, as shown in Table 1, the improved YOLO achieves a 6.1% improvement in accuracy over the original YOLOv11. While the improved YOLOv11's accuracy is lower than the 96.17% of the LA-YOLO-LLL model, the disclosed embodiments, which consider the F1 and mAP50 (mean Average Precision at IoU = 0.5, which calculates AP (Average Precision)—a prediction is considered a "correct detection" only when the IoU between the predicted box and the ground-truth box is ≥ 0.5) parameters—achieve precision and recall rates that are 3.17% and 3.7% higher, respectively, than the LA-YOLO-LLL model. In terms of recall, the improved YOLOv11, or landslide detection model, provided by the disclosed embodiments, surpasses the second-best performing YOLOv8 by 3.4%, reaching 99.5%, significantly exceeding all other models. In addition, the improved YOLOv11, i.e., the landslide detection model provided by the embodiments of the present disclosure, has a mAP50 that is 99.1%, 3.4%, and 3% higher than that of YOLOv8 and YOLOv11, respectively, and a mAP@50-95 (mean Average Precision at IoU=0.5:0.95, i.e., the average mAP value when the IoU threshold is calculated from 0.5 to 0.95 (step size 0.05)) that is 3.2% and 3.8% higher than that of YOLOv8 and YOLOv11, respectively.

[0094] As can be seen, the improved YOLOv11, or landslide detection model, exhibits excellent performance and stability for a variety of challenging landslide images. For example, targets smaller than 32*32 pixels are defined as small targets. Due to feature limitations, these targets are difficult for conventional models to learn, but the improved YOLOv11 provided by the present disclosure can still accurately and stably detect them. Furthermore, objects vary in shape and size, and some may even have very small, large, or extreme shapes (such as slender, narrow, or tall). This poses significant challenges for accurate target identification and positioning. However, the improved YOLOv11 provided by the present disclosure still demonstrates excellent stability in this type of detection. Furthermore, due to factors such as vegetation occlusion and surface differentiation, the boundaries between some targets and the background are unclear, and the overall feature representation of the target is weak, which can easily lead to missed detections and false detections by the model. However, the improved YOLOv11 provided by the present disclosure still performs well in such situations. Therefore, when performing landslide detection on remote sensing images, the improved YOLOv11 provided by the present disclosure can efficiently and accurately detect the landslide type in the remote sensing image being detected.

[0095] The exemplary embodiment of the present disclosure also provides a landslide detection device. Figure 8 As shown, the landslide detection device 800 includes the following program units: A determination unit 801 is used to determine a remote sensing image to be detected; The detection unit 802 is used to perform feature extraction on the remote sensing image to be detected through the backbone network of the landslide detection model to obtain multiple initial feature maps of different sizes; and to fuse the multiple initial feature maps of different sizes through the neck network of the landslide detection model to obtain multiple fused feature maps of different sizes; to perform feature processing on the multiple fused feature maps of different sizes through the receptive field attention submodule of the landslide detection model to obtain target receptive field features; and to detect and process the target receptive field features through the detection network of the landslide detection model to obtain a landslide detection frame and a landslide category label.

[0096] In a possible implementation, the detection unit 802 is specifically configured to: Inputting the multiple fused feature maps of different sizes into the first sub-branch of the receptive field attention sub-module to obtain a first feature matrix; Inputting the multiple fused feature maps of different sizes into the second sub-branch of the receptive field attention sub-module to obtain a second feature matrix; Input the second feature matrix into the third sub-branch of the receptive field attention sub-module to obtain a third feature matrix; Determine target receptive field features based on the first feature matrix and the third feature matrix.

[0097] In a possible implementation, the detection unit 802 is specifically configured to: The multiple fused feature maps of different sizes are input into the nonlinear activation function of the receptive field attention sub-module to obtain the channel attention vector of the original feature corresponding to the fused feature map, and the channel attention vector is used as the first feature matrix.

[0098] In a possible implementation, the detection unit 802 is specifically configured to: Performing global average pooling and one-dimensional convolution processing on the multiple fused feature maps of different sizes to obtain a first sub-feature matrix; The first sub-feature matrix is processed using convolution kernels of different sizes and dilated convolution to obtain a second feature matrix.

[0099] In a possible implementation, the detection unit 802 is specifically configured to:

[0100] in, X out Used to characterize the target receptive field characteristics; among them, , X FUsed to characterize the second characteristic matrix, X Used to characterize the feature matrix corresponding to the fused feature map; X a Used to characterize the third characteristic matrix; Used to characterize the first characteristic matrix; Re LU ( X a ) is used to represent the application of the ReLU activation function to each element in the third feature matrix.

[0101] In a possible implementation, the device further includes a training unit configured to: Collecting a remote sensing image data set and dividing the remote sensing image data set into a training set, a validation set, and a test set; Training a preset landslide detection model according to the training set to obtain multiple prediction information; Determining comparison information between the plurality of prediction information and the labeled information in the training set through a loss function, and adjusting the preset landslide detection model according to the comparison information to obtain an adjusted preset landslide detection model; Adjusting the hyperparameters in the adjusted preset landslide detection model according to the validation set, and obtaining an updated preset landslide detection model when it is determined that the validation set loss reaches a first preset condition; The updated preset landslide detection model is evaluated according to the test set, and when it is determined that the evaluation result meets the second preset condition, the landslide detection model is obtained.

[0102] In one possible implementation, the loss function is determined based on the following formula:

[0103] in, N 1 is used to represent the prediction information, N 2 is used to represent the annotation information, C 2 is used to represent a constant related to the training set, Used to characterize the second-order Wasserstein distance between the predicted information and the labeled information; NWD ( N 1 , N 2 ) is used to characterize the similarity between the predicted information and the labeled information.

[0104] The specific details of each part of the above-mentioned device have been described in detail in the implementation method part. The undisclosed details can be found in the implementation method part, so they will not be repeated here.

[0105] It should be noted that although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the exemplary embodiments of the present disclosure, the features and functions of two or more modules or units described above can be concretized in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided into multiple modules or units to be concretized.

[0106] The exemplary embodiments of the present disclosure further provide a computer program product, which includes a computer program, and when the computer program is executed by a processor, implements the above-mentioned landslide detection method.

[0107] In one embodiment, a computer program product may be a tangible product containing a computer program, such as a computer-readable storage medium storing the computer program. The computer-readable storage medium may be a storage medium based on electrical, magnetic, optical, electromagnetic, infrared, or other signals, including but not limited to random access memory (RAM), read-only memory (ROM), magnetic tape, floppy disk, flash memory (Flash), hard disk drive (HDD), solid-state drive (SSD), and the like. Exemplarily, the computer program product may be implemented as a non-volatile storage medium storing the computer program, such as a read-only memory (ROM) or NAND flash memory.

[0108] In one embodiment, the computer program product may be an intangible product containing a computer program. For example, the computer program product may be implemented as a virtual digital product, such as a digital file such as an executable file or installation package storing the computer program.

[0109] The code of a computer program can be written in one or more programming languages, such as C, Java, C++, and the like. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving a remote computing device, the remote computing device can be connected to the user's computing device via any type of network, such as a local area network (LAN) or wide area network (WAN), or can be connected to an external computing device (e.g., via an internet connection provided by a carrier).

[0110] Computer programs can be carried or transmitted via electrical, magnetic, optical, electromagnetic, infrared, or other signals. Electronic devices can convert signals carrying computer programs into digital signals, thereby executing the computer programs. When the computer program is executed on an electronic device, its code causes the electronic device to execute (more specifically, the processor of the electronic device to execute) the method steps of various exemplary embodiments of the present disclosure, such as the aforementioned landslide detection method, which includes the following steps: Step 501: Determine a remote sensing image to be detected; Step 502: Perform feature extraction on the remote sensing image to be detected using the backbone network of a landslide detection model to obtain multiple initial feature maps of varying sizes; and, Perform fusion processing on the multiple initial feature maps of varying sizes using the neck network of the landslide detection model to obtain multiple fused feature maps of varying sizes; Step 504: Perform feature processing on the multiple fused feature maps of varying sizes using the receptive field attention submodule of the landslide detection model to obtain target receptive field features; Step 505: Perform detection processing on the target receptive field features using the detection network of the landslide detection model to obtain a landslide detection box and a landslide category label.

[0111] By implementing the above method steps through a computer program, a remote sensing image to be detected can be determined, and then the backbone network of the landslide detection model is used to extract features from the remote sensing image to be detected, thereby obtaining multiple initial feature maps of different sizes; and the neck network of the landslide detection model is used to fuse the multiple initial feature maps of different sizes, thereby obtaining multiple fused feature maps of different sizes; the receptive field attention submodule of the landslide detection model is used to perform feature processing on the multiple fused feature maps of different sizes, thereby obtaining target receptive field features; and the detection network of the landslide detection model is used to detect and process the target receptive field features, thereby obtaining a landslide detection frame and a landslide category label. It can be seen that the present disclosure adopts a landslide detection model, which not only can automatically detect the remote sensing image to be detected, thereby avoiding errors caused by manual analysis and improving detection accuracy, but also combines the receptive field attention submodule with the features determined by deep mining of the remote sensing image to be detected to perform landslide detection on the remote sensing image to be detected, thereby further improving the accuracy of detection of the remote sensing image to be detected.

[0112] The exemplary embodiments of the present disclosure further provide an electronic device, which may include a processor and a memory. The memory stores executable instructions for the processor, such as a computer program. The processor executes the executable instructions to perform the method steps of various exemplary embodiments of the present disclosure.

[0113] Reference below Figure 9 , the electronic device is exemplarily described in the form of a general-purpose computing device. It should be understood that Figure 9 The electronic device 120 shown is merely an example and should not limit the functionality and scope of use of the embodiments of the present disclosure.

[0114] like Figure 9 As shown, the electronic device 120 may include a processor 910 , a memory 920 , a bus 930 , an I / O (input / output) interface 940 , and a network adapter 950 .

[0115] The memory 920 may include volatile memory, such as RAM 921 and cache unit 922, and may also include non-volatile memory, such as ROM 923. The memory 920 may also include one or more program modules 924. Such program modules 924 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. For example, the program modules 924 may include the various units in the aforementioned apparatus.

[0116] The processor 910 may include one or more processing units. For example, the processor 910 may include an AP (Application Processor), a modem processor, a GPU (Graphics Processing Unit), an ISP (Image Signal Processor), a controller, an encoder, a decoder, a DSP (Digital Signal Processor), a baseband processor and / or an NPU (Neural-Network Processing Unit), etc.

[0117] The processor 910 can be used to execute executable instructions stored in the memory 920, such as the above-mentioned landslide detection method, which includes the following steps: Step 501: Determine the remote sensing image to be detected; Step 502: Perform feature extraction on the remote sensing image to be detected through the backbone network of the landslide detection model to obtain multiple initial feature maps of different sizes; and, Perform fusion processing on the multiple initial feature maps of different sizes through the neck network of the landslide detection model to obtain multiple fused feature maps of different sizes; Step 504: Perform feature processing on the multiple fused feature maps of different sizes through the receptive field attention submodule of the landslide detection model to obtain target receptive field features; Step 505: Perform detection processing on the target receptive field features through the detection network of the landslide detection model to obtain a landslide detection frame and a landslide category label.

[0118] By executing the above method steps by the processor 910, a remote sensing image to be detected can be determined, and then the backbone network of the landslide detection model is used to extract features from the remote sensing image to be detected to obtain multiple initial feature maps of different sizes; and the neck network of the landslide detection model is used to fuse the multiple initial feature maps of different sizes to obtain multiple fused feature maps of different sizes; the receptive field attention submodule of the landslide detection model is used to perform feature processing on the multiple fused feature maps of different sizes to obtain target receptive field features; and the detection network of the landslide detection model is used to detect and process the target receptive field features to obtain a landslide detection frame and a landslide category label. It can be seen that the present disclosure adopts a landslide detection model, which can not only automatically detect the remote sensing image to be detected, avoiding errors caused by manual analysis and improving detection accuracy, but also combines the receptive field attention submodule to perform landslide detection on the remote sensing image to be detected based on the features determined by deep mining of the remote sensing image to be detected, further improving the accuracy of detection of the remote sensing image to be detected.

[0119] The bus 930 is used to realize the connection between different components of the electronic device 120 and may include a data bus, an address bus, and a control bus.

[0120] The electronic device 120 can communicate with one or more external devices 1000 (eg, a keyboard, a mouse, an external controller, etc.) through the I / O interface 940 .

[0121] The electronic device 120 can communicate with one or more networks via the network adapter 950. For example, the network adapter 950 can provide mobile communication solutions such as 3G / 4G / 5G, or wireless communication solutions such as wireless LAN, Bluetooth, and near-field communication. The network adapter 950 can communicate with other modules of the electronic device 120 via the bus 930.

[0122] although Figure 9 Not shown, other hardware and / or software modules may also be provided in the electronic device 120, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0123] As can be seen from the above, the technical solutions of the present disclosure can be implemented as methods, devices, systems, computer program products, storage media, electronic devices, etc. Those skilled in the art will appreciate that various aspects of the present disclosure can be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or an implementation that combines hardware and software aspects, such as "circuits," "modules," or "systems," respectively.

[0124] It should be understood that the present disclosure is not limited to the specific method steps or structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from the scope thereof. Those skilled in the art will easily think of other embodiments based on the specific embodiments provided by the present disclosure. Therefore, the specific embodiments provided by the present disclosure are merely exemplary, and the scope and spirit of the present disclosure are indicated by the claims, which should cover any variations, uses or adaptive changes of the present disclosure, which follow the general principles of the present disclosure and include common knowledge or customary technical means in the field of the present technology that are not disclosed in the present disclosure.

Claims

1. A landslide detection method, characterized in that: The method comprises: Determine the remote sensing image to be detected; Performing feature extraction on the remote sensing image to be detected through a backbone network of the landslide detection model to obtain a plurality of initial feature maps of different sizes; and performing fusion processing on the plurality of initial feature maps of different sizes through a neck network of the landslide detection model to obtain a plurality of fused feature maps of different sizes; Performing feature processing on the multiple fused feature maps of different sizes through the receptive field attention submodule of the landslide detection model to obtain target receptive field features; The target receptive field features are detected and processed by the detection network of the landslide detection model to obtain a landslide detection frame and a landslide category label.

2. The method according to claim 1, characterized in that The receptive field attention submodule of the landslide detection model performs feature processing on the multiple fused feature maps of different sizes to obtain target receptive field features, including: Inputting the multiple fused feature maps of different sizes into the first sub-branch of the receptive field attention sub-module to obtain a first feature matrix; Inputting the multiple fused feature maps of different sizes into the second sub-branch of the receptive field attention sub-module to obtain a second feature matrix; Input the second feature matrix into the third sub-branch of the receptive field attention sub-module to obtain a third feature matrix; Determine target receptive field features based on the first feature matrix and the third feature matrix.

3. The method according to claim 2, characterized in that Inputting the multiple fused feature maps of different sizes into the first sub-branch of the receptive field attention sub-module to obtain a first feature matrix, including: The multiple fused feature maps of different sizes are input into the nonlinear activation function of the receptive field attention sub-module to obtain the channel attention vector of the original feature corresponding to the fused feature map, and the channel attention vector is used as the first feature matrix.

4. The method according to claim 2, characterized in that Inputting the multiple fused feature maps of different sizes into the second sub-branch of the receptive field attention sub-module to obtain a second feature matrix, including: Performing global average pooling and one-dimensional convolution processing on the multiple fused feature maps of different sizes to obtain a first sub-feature matrix; The first sub-feature matrix is processed using convolution kernels of different sizes and dilated convolution to obtain a second feature matrix.

5. The method according to claim 2, characterized in that Determining target receptive field features according to the first feature matrix and the third feature matrix includes: in, X out Used to characterize the target receptive field characteristics; among them, , X F Used to characterize the second characteristic matrix, X Used to characterize the feature matrix corresponding to the fused feature map; X a Used to characterize the third characteristic matrix; Used to characterize the first characteristic matrix; Re LU ( X a ) is used to represent the application of the ReLU activation function to each element in the third feature matrix.

6. The method according to any one of claims 1 to 5, characterized in that: The landslide detection model is trained based on the following method: Collecting a remote sensing image data set and dividing the remote sensing image data set into a training set, a validation set, and a test set; Training a preset landslide detection model according to the training set to obtain multiple prediction information; Determining comparison information between the plurality of prediction information and the labeled information in the training set through a loss function, and adjusting the preset landslide detection model according to the comparison information to obtain an adjusted preset landslide detection model; Adjusting the hyperparameters in the adjusted preset landslide detection model according to the validation set, and obtaining an updated preset landslide detection model when it is determined that the validation set loss reaches a first preset condition; The updated preset landslide detection model is evaluated according to the test set, and when it is determined that the evaluation result meets the second preset condition, the landslide detection model is obtained.

7. The method according to claim 6, characterized in that The loss function is determined based on the following formula: in, N 1 is used to represent the prediction information, N 2 is used to represent the annotation information, C 2 is used to represent a constant related to the training set, Used to characterize the second-order Wasserstein distance between the predicted information and the labeled information; NWD ( N 1 , N 2 ) is used to characterize the similarity between the predicted information and the labeled information.

8. A landslide detection device, characterized in that: The device comprises: A determination unit, used for determining a remote sensing image to be detected; The detection unit is used to perform feature extraction on the remote sensing image to be detected through the backbone network of the landslide detection model to obtain multiple initial feature maps of different sizes; and to fuse the multiple initial feature maps of different sizes through the neck network of the landslide detection model to obtain multiple fused feature maps of different sizes; to perform feature processing on the multiple fused feature maps of different sizes through the receptive field attention submodule of the landslide detection model to obtain target receptive field features; and to detect and process the target receptive field features through the detection network of the landslide detection model to obtain a landslide detection frame and a landslide category label.

9. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

10. An electronic device, characterized in that: include: processor; a memory for storing executable instructions of the processor; The processor is configured to perform the method according to any one of claims 1 to 7 by executing the executable instructions.

Images