Remote sensing image target detection method based on hyperbolic space mapping and electronic equipment

Abstract

The remote sensing image target detection method based on hyperbolic space mapping comprises the following steps: S1, extracting features of a remote sensing image to obtain a hierarchical feature map; s2, fusing the hierarchy feature maps of the same type through a feature fusion network to obtain a first multi-hierarchy feature map; s3, projecting channel information of the first multi-level feature map to a hyperbolic space by using a hyperbolic space mapping network to obtain a second multi-level feature map; s4, splicing the first multi-level feature map and the second multi-level feature map to obtain a fused feature map; s5, constructing a feature detection head, detecting a fusion feature map, and calculating classification loss and position prediction loss; s6, repeating S1 to S5, and training a remote sensing image target detection model; and S7, detecting a remote sensing image by using the remote sensing image target detection model obtained in S6. According to the method, by enhancing the information dimension of the prediction feature map, the model is helped to understand data distribution and learn more abstract feature representation, so that the generalization ability of a target detection task is improved.

Description

Technical Field

[0001] The invention relates to the technical field of remote sensing satellite data processing, and in particular to a remote sensing image target detection method and electronic equipment based on hyperbolic space mapping. Background Art

[0002] With the continuous development of remote sensing technology and the continuous improvement of satellite resolution, the number and resolution of remote sensing images are also increasing, and the number of remote sensing images acquired is growing exponentially. Remote sensing image target detection refers to the process of automatically detecting targets of interest from remote sensing images acquired from long distances such as satellites or aerial vehicles. This technology is widely used in many fields, such as agriculture, forestry, urban planning, natural resource management, and environmental monitoring. Remote sensing images have the detection difficulties of large width, complex imaging background, and a large number of small targets. Therefore, how to efficiently detect targets from a large number of remote sensing images has become an important research direction.

[0003] Currently, the commonly used remote sensing image target detection technologies include traditional feature engineering-based methods and deep learning-based methods. Traditional feature engineering-based methods require manual extraction of features in images and use machine learning algorithms to classify these features. This method requires manual design of feature extraction rules and is difficult to process complex remote sensing image data. The deep learning-based method uses a deep convolutional neural network (CNN) to automatically learn features in images and achieve end-to-end target detection. Therefore, it has high accuracy and robustness, and does not require manual feature extraction. Therefore, it is widely used in the field of remote sensing image target detection.

[0004] With the continuous development of deep learning technology, deep learning-based methods are also constantly improving and optimizing, such as introducing attention mechanisms, multi-task learning and other technologies to improve the accuracy of detection. However, most of the current remote sensing image target detectors based on deep learning simply decouple the information inside the feature map based on the calculation method of Euclidean space. This type of strategy cannot effectively capture the high-dimensional spatial geometric metric information between feature map information, and the high-dimensional spatial geometric metric information between targets (such as the structural properties of global features and local features, relative position relationships, etc.) is very important for detection tasks. Therefore, it is difficult to effectively utilize the differences and complementarities between feature data, resulting in poor results of target remote sensing image target detection models. The above problems have hindered the existing remote sensing image target detection tasks from achieving better performance. Summary of the invention

[0005] In order to solve the technical problems existing in the above-mentioned prior art, the purpose of the present invention is to provide a remote sensing image target detection method and electronic equipment based on hyperbolic space mapping, which helps the target remote sensing image target detection model to capture the geometric metric information of high-dimensional feature data in the high-dimensional hyperbolic space, and use this information to assist prediction in the feature decoupling stage, enhance the information dimension of the predicted feature map, help the model better understand the distribution of the data, learn more abstract feature representations, and improve the generalization ability of target detection tasks.

[0006] To achieve the above-mentioned object of the invention, the present invention provides a remote sensing image target detection method based on hyperbolic space mapping, comprising the following steps:

[0007] Step S1, extracting features of an input remote sensing image through a feature extraction backbone network, and obtaining a plurality of hierarchical feature maps corresponding to the convolutional levels of the feature extraction backbone network;

[0008] Step S2: using a feature fusion network, according to the type of the hierarchical feature map, the hierarchical feature maps of the same type are fused to obtain a first plurality of hierarchical feature maps of the remote sensing image target, wherein the types of the hierarchical feature maps include at least a deep feature map, a middle feature map, and a shallow feature map;

[0009] Step S3, using a hyperbolic space mapping network to project the channel information of the plurality of the first multi-level feature maps into a hyperbolic space, to obtain a plurality of second multi-level feature maps, and to obtain geometric metric information of the feature information in a high-dimensional hyperbolic space;

[0010] Step S4, concatenating the first multi-level feature map and the second multi-level feature map to obtain a plurality of fused feature maps;

[0011] Step S5: construct a feature detection head, detect the fused feature map, and calculate the classification loss and the position prediction loss;

[0012] Step S6, repeating steps S1 to S5 to train a remote sensing image target detection model, wherein the remote sensing image target detection model includes the feature extraction backbone network, the feature fusion network, the hyperbolic space mapping network and the feature detection head;

[0013] Step S7: Detect the remote sensing image using the remote sensing image target detection model obtained in step S6.

[0014] According to a technical solution of the present invention, in the step S1, it specifically includes:

[0015] Step S11, constructing the feature extraction backbone network;

[0016] Step S12, performing a preprocessing operation on the input remote sensing image; the preprocessing operation includes image cropping, image flipping, projection transformation, mosaic enhancement and / or image filling;

[0017] Step S13: performing feature extraction operation on the preprocessed remote sensing image through the feature extraction backbone network to obtain the hierarchical feature map.

[0018] According to a technical solution of the present invention, in the step S2, it specifically includes:

[0019] Step S21, constructing the feature fusion network;

[0020] Step S22: receiving the hierarchical feature map through the feature fusion network, and performing a fusion operation on the hierarchical feature maps of the same type according to the type of the hierarchical feature map to obtain the first multi-layer feature map; the fusion operation includes upsampling, lateral connection and / or dimensional splicing.

[0021] According to a technical solution of the present invention, in the step S22, it specifically includes:

[0022] Step S221, traversing all the hierarchical feature maps, and performing channel number consistency processing on the hierarchical feature maps;

[0023] Step S222: perform a size upsampling operation on the hierarchical feature map after the channel number consistency processing, adjust the size of the shallow feature map to the size of the middle feature map, and then perform dimensional splicing or channel addition to obtain the first multi-level feature map.

[0024] According to a technical solution of the present invention, in the step S3, the hyperbolic space mapping network maps the first multi-level feature map vector by a hyperbolic space mapping algorithm, performs pixel-level data traversal on the channel information of the first multi-level feature map, extracts the feature vector of the 1*1 channel number according to the pixel cycle, obtains the corresponding number of feature points, and then performs hyperbolic space mapping processing on the obtained feature points to obtain the feature vector;

[0025] The hyperbolic space mapping algorithm is modeled as a Möbius transform or a Poincare disk.

[0026] According to a technical solution of the present invention, in step S3, the channel information of the feature map is projected into the hyperbolic space using the Mobius transform, which specifically includes:

[0027] Step S31: Perform vector representation on the first multi-level feature map. For any first multi-level feature map, define F∈R H×W×D is the first multi-level feature map from the feature fusion network, with a size of H×W and a number of channels of D;

[0028] F = [x1, x2, ..., x i , …, x N ]

[0029] Among them, [x1, x2, …, xi ， …, x N ] represents the sequence data obtained after the first multi-level feature map F is stretched, N = H × W, x i ∈R N*D Represents the two-dimensional vector corresponding to the feature point at the i-th position in the first multi-level feature map, i∈[1, 2, …, N];

[0030] Step S32: Map the channel information of the plurality of the first multi-level feature maps in the Euclidean distance space to the Riemann sphere space using the Mobius transformation method to obtain data structure property information in the surface space. The overall overview of the Mobius transformation method is as follows:

[0031] Define the extended complex plane The Mobius transform of , where C is a complex plane, ∞ is a point at infinity, and for the two-dimensional vector x corresponding to the feature point at the i-th position in the first multi-level feature map F i =[N i , D i ], after the Möbius transformation τ(·), we get:

[0032]

[0033] Where S is a reversible matrix used to represent the projection transformation parameters a, b, c, d:

[0034]

[0035] For the two-dimensional vector x corresponding to the feature point at the i-th position in the first multi-level feature map F i , the universal Möbius transformation T(x i ) is expressed as follows:

[0036]

[0037] The Möbius transform remains unchanged by scaling the projection transformation parameters. The Möbius transform has at most two fixed points, and the fixed point γ can be obtained by solving:

[0038]

[0039] After the Möbius transformation, the feature matrix F of the similar shape of the first multi-level feature map F can be obtained. Mobius , its internal data is characterized by the feature mapping of the feature data of the first multi-level feature map F in the Riemann sphere space, that is:

[0040] F Mobius =[x Mobius，1 , x Mobius，2 , …, x, Mobius，i , …, x Mobius，M ]

[0041] where F Mobius is the feature matrix of the second multi-level feature map obtained by transforming the first multi-level feature map F, [x, Mobius，i ], indicating [x i ] corresponds to the transformation structure, M represents the sequence number of the feature data obtained by the transformation.

[0042] According to a technical solution of the present invention, the step S4 specifically includes:

[0043] Step S41, reshaping the plurality of second multi-level feature maps obtained in step S3;

[0044] Step S42, performing feature normalization on the plurality of second multi-level feature maps obtained after feature reshaping in step S41, and normalizing them to a range of 0 to 255;

[0045] Step S43, using the normalized second multi-level feature maps to fuse with the first multi-level feature maps in step S2 to obtain a plurality of fused feature maps; wherein the fusion method includes an addition operation, a feature concatenation operation and / or a dual fusion operation;

[0046] Step S44, filtering the multiple fused feature maps obtained in step S43 through an attention mechanism to obtain an optimal channel information representation;

[0047] Step S45, using the fused feature map screened out in step S44 to transmit and send to the feature detection head.

[0048] According to a technical solution of the present invention, in the step S41, it specifically includes:

[0049] The feature matrix F of the second multi-level feature map obtained after step S33 is reshaped by the reshape operation Mobius Process it and normalize it to the first multi-level feature map F∈R H×W×D within the domain.

[0050] According to a technical solution of the present invention, in step S6, the simulated prediction result of the remote sensing image target detection model is obtained, and the difference between the simulated prediction result and the real annotation of the input remote sensing image is used as a judgment basis to control the training.

[0051] According to one aspect of the present invention, there is provided an electronic device, comprising: one or more processors, one or more memories, and one or more computer programs; wherein the processor is connected to the memory, the one or more computer programs are stored in the memory, and when the electronic device is running, the processor executes the one or more computer programs stored in the memory, so that the electronic device performs the above-mentioned remote sensing image target detection method based on hyperbolic space mapping.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] The invention proposes a remote sensing image target detection method based on hyperbolic space mapping. When training a remote sensing image target detection model based on a convolutional neural network, the features of an input remote sensing image are extracted through a feature extraction backbone network to obtain a hierarchical feature map, thereby improving the model's ability to classify and locate interested targets in the remote sensing image. Then, a fused shallow feature map, a middle feature map, and a deep feature map are obtained through a feature fusion network. The remote sensing image target detection method based on hyperbolic space mapping projects the shallow feature map, the middle feature map, and the deep feature map generated in the feature fusion network to a hyperbolic space to obtain geometric metric information of feature information in a high-dimensional hyperbolic space, thereby helping the target remote sensing image target detection model to capture the spatial geometric metric information of high-dimensional feature data in the feature fusion stage, thereby enhancing the information dimension of the original prediction feature map, helping the model to better understand the distribution of data, learn more abstract feature representations, and enhancing the information dimension of the original prediction feature map, thereby improving the generalization ability of target detection tasks.

[0054] Furthermore, the remote sensing image target detection model can obtain multi-scale information of the target during the detection process, and obtain richer and more accurate target information, thereby improving the model's ability to classify and locate targets of interest in remote sensing images, and improving the accuracy of remote sensing image target positioning, which is of great significance for high-resolution remote sensing image rotating frame target detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0056] Figure 1 Schematically showing a flow chart of a remote sensing image target detection method based on hyperbolic space mapping according to an embodiment of the present invention;

[0057] Figure 2A schematic diagram schematically shows the structure of a remote sensing image target detection model according to another embodiment of the present invention;

[0058] Figure 3 Schematically showing an overall flow chart of a feature fusion target detection algorithm based on hyperbolic space mapping according to another embodiment of the present invention;

[0059] Figure 4 The model algorithm flow chart of the remote sensing image target detection method based on hyperbolic space mapping in another embodiment of the present invention is schematically shown. DETAILED DESCRIPTION

[0060] The description of the embodiments of this specification should be combined with the corresponding drawings, which should be considered as part of the complete specification. In the drawings, the shape or thickness of the embodiments may be enlarged and indicated for simplification or convenience. Furthermore, the parts of each structure in the drawings will be described separately. It is worth noting that the elements not shown in the drawings or not described in words are in a form known to ordinary technicians in the relevant technical field.

[0061] The description of the embodiments herein and any reference to directions and orientations are only for the convenience of description and are not to be construed as any limitation on the scope of protection of the present invention. The following description of the preferred embodiments may involve combinations of features, which may exist independently or in combination, and the present invention is not particularly limited to the preferred embodiments. The scope of the present invention is defined by the claims.

[0062] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0063] like Figures 1 to 3 As shown, the remote sensing image target detection method based on hyperbolic space mapping of this embodiment includes the following steps:

[0064] Step S1, extracting features of the input remote sensing image through a feature extraction backbone network, and obtaining a plurality of hierarchical feature maps corresponding to the convolutional levels of the feature extraction backbone network;

[0065] Step S1 specifically includes:

[0066] Step S11, constructing a feature extraction backbone network;

[0067] Step S12, performing preprocessing operations on the input remote sensing image; the preprocessing operations include image cropping, image flipping, projection transformation, mosaic enhancement and / or image filling;

[0068] Step S13: Perform feature extraction on the preprocessed remote sensing image through a feature extraction backbone network to obtain a hierarchical feature map.

[0069] By performing operations such as cropping and flipping on remote sensing images, it is helpful to enhance the robustness, universality and generalization ability of the remote sensing image target detection model algorithm.

[0070] In the process of training remote sensing image target detection models, the training data set used is one of the important factors affecting the model performance. By manually and algorithmically screening and cleaning the data set and removing data that does not meet the requirements or contains erroneous information, the interference of invalid data can be reduced and the accuracy and reliability of the model can be improved.

[0071] Step S2: using a feature fusion network, according to the type of the hierarchical feature map, the hierarchical feature maps of the same type are fused to obtain a first plurality of hierarchical feature maps of the remote sensing image target, wherein the types of the hierarchical feature maps include at least a deep feature map, a middle feature map, and a shallow feature map;

[0072] In step S2, it specifically includes:

[0073] Step S21, constructing a feature fusion network;

[0074] Step S22, receiving the hierarchical feature map through the feature fusion network, and performing a fusion operation on the hierarchical feature maps of the same type according to the type of the hierarchical feature map to obtain a first multi-layer hierarchical feature map; the fusion operation includes upsampling, lateral connection and / or dimensional splicing.

[0075] When training a remote sensing image target detection model based on a convolutional neural network, constructing a feature fusion network through a feature pyramid can enable the model to obtain multi-dimensional information of the target, thereby improving the model's ability to classify and locate targets of interest in remote sensing images.

[0076] Step S22 retains the traditional feature fusion method, which facilitates the subsequent use of hyperbolic space mapping to assist the feature fusion process, so that more valuable information is retained or enhanced. Step S22 specifically includes:

[0077] Step S221, traverse all hierarchical feature maps, and use 1*1 convolution to process the hierarchical feature maps to make the number of channels consistent;

[0078] Step S222: perform a size upsampling operation on the hierarchical feature map after the channel number is processed to be consistent, adjust the size of the shallow feature map to the size of the middle feature map, and then perform dimensional splicing or channel addition to obtain the first multi-level feature map.

[0079] Step S3, using a hyperbolic space mapping network to project the channel information of the plurality of first multi-level feature maps into the hyperbolic space, to obtain a plurality of second multi-level feature maps, and to obtain geometric metric information of the feature information in the high-dimensional hyperbolic space;

[0080] In step S3, the hyperbolic space mapping network maps the first multi-level feature map vector by a hyperbolic space mapping algorithm, performs pixel-level data traversal on the channel information of the first multi-level feature map, extracts the feature vector of the 1*1 channel number according to the pixel cycle, obtains the corresponding number of feature points, and then performs hyperbolic space mapping processing on the obtained feature points to obtain the feature vector;

[0081] Hyperbolic space mapping algorithms model the Möbius transformation or the Poincare disk.

[0082] In step S3, the channel information of the feature map is projected into the hyperbolic space using the Möbius transform, which specifically includes:

[0083] Step S31: vectorize the first multi-level feature map. For any first multi-level feature map, define F∈R H×W×D is the first multi-level feature map from the feature fusion network, with a size of H×W and a number of channels of D;

[0084] F=[x1,x2,…,x i ,…,x N ]

[0085] Among them, [x1,x2,…,x i ,…,x N ] represents the sequence data obtained after the first multi-level feature map F is stretched, N = H × W, x i ∈R N*D Represents the two-dimensional vector corresponding to the feature point at the i-th position in the first multi-level feature map, i∈[1,2,…,N];

[0086] Step S32: Map the channel information of the first multi-level feature maps in the Euclidean distance space to the Riemann sphere space using the Mobius transformation method to obtain the data structure property information in the surface space. The overall overview of the Mobius transformation method is as follows:

[0087] Define the extended complex plane The Mobius transform of , where С is the complex plane, ∞ is the point at infinity, and for the two-dimensional vector x corresponding to the feature point at the i-th position in the first multi-level feature map F i =[N i ,D i ], after the Möbius transformation τ(·), we can get:

[0088]

[0089] Where S is a reversible matrix used to represent the projection transformation parameters a, b, c, d:

[0090]

[0091] For the two-dimensional vector x corresponding to the feature point at the i-th position in the first multi-level feature map F i , the universal Möbius transformation τ(x i ) can be expressed as follows:

[0092]

[0093] The Möbius transform remains unchanged through the scaling of the projection transformation parameters. The Möbius transform has at most two fixed points (i.e., the identity mapping case), but the identity mapping is not considered in practical applications, so there is only one, so the fixed point γ can be obtained by solving:

[0094]

[0095] After the Möbius transformation, we can get the feature matrix F of the similar shape of the first multi-level feature map F Mobius , its internal data is characterized by the feature mapping of the feature data of the first multi-level feature map F in the Riemann sphere space, that is:

[0096] F Mobius =[x Mobius,1 ,x Mobius,2 ,…,x ,Mobius,i ,…,x Mobius,M ]

[0097] where F Mobius is the feature matrix of the second multi-level feature map obtained by transforming the first multi-level feature map F, [x ,Mobius,i ] means [x i ] corresponds to the transformation structure, M represents the sequence number of the feature data obtained by the transformation;

[0098] Therefore, the above method is used to obtain multiple second-layer multi-level feature maps after Möbius transformation, which can be involved in the feature fusion process to utilize the geometric metric representation of feature information in high-dimensional hyperbolic space to assist model prediction. Feature fusion methods include feature concatenation, feature summation and / or skip connection.

[0099] In this embodiment, hyperbolic space mapping is used to perform channel information mapping processing on the shallow, middle and deep feature maps generated in the feature fusion network to obtain geometric measurement information of the feature information in the high-dimensional hyperbolic space, and then the multiple feature maps obtained based on the hyperbolic space mapping are fused with the original feature map, so as to help the target remote sensing image target detection model to capture the spatial geometric measurement information of the high-dimensional feature data in the feature fusion stage, and to use the spatial geometric measurement information of the high-dimensional feature data to assist prediction in the feature decoupling stage, thereby enhancing the information dimension of the original prediction feature map, helping the model to better understand the distribution of the data, learn more abstract feature representations, and enhance the information dimension of the original prediction feature map, thereby improving the generalization ability of the target detection task.

[0100] Step S4, concatenating the first multi-level feature map and the second multi-level feature map to obtain a plurality of fused feature maps;

[0101] Step S4 specifically includes:

[0102] Step S41, reshaping the plurality of second multi-level feature maps obtained in step S3 to meet subsequent fusion requirements;

[0103] The feature matrix F of the second multi-level feature map obtained after step S33 is reshaped by the reshape operation Mobius Process it and normalize it to the first multi-level feature map F∈R H×W×D domain to obtain the feature map size that meets the subsequent fusion requirements.

[0104] Step S42, performing feature normalization on the plurality of second multi-level feature maps obtained after feature reshaping in step S41 to obtain data representation within a suitable channel range, that is, normalizing to a range of 0 to 255;

[0105] Step S43, fusing the normalized second multi-level feature maps with the first multi-level feature maps in step S2 to obtain a plurality of fused feature maps; wherein the fusion method includes an addition operation, a feature concatenation operation and / or a dual fusion operation;

[0106] Step S44, filtering the multiple fusion feature maps obtained in step S43 through the attention mechanism to obtain the optimal channel information representation;

[0107] Step S45, using the fused feature map filtered out in step S44 for transmission, and sending it to the feature detection head for subsequent feature decoupling.

[0108] Step S5: construct a feature detection head, detect the fused feature map, and calculate the classification loss and position prediction loss.

[0109] Step S6, repeating steps S1 to S5 to train the remote sensing image target detection model, the remote sensing image target detection model includes a feature extraction backbone network, a feature fusion network, a hyperbolic space mapping network and a feature detection head.

[0110] The overall training process of the remote sensing image target detection model in this embodiment is as follows: Figure 3 As shown, first, the remote sensing image and its real annotation information are read, and the preprocessing steps are performed on the remote sensing image. After completing the construction of the backbone feature extraction network, this embodiment adopts a feature fusion network architecture based on hyperbolic space mapping enhancement, which is different from the traditional feature fusion network. The network first receives a multi-layer feature map, and then performs a feature transformation process based on hyperbolic space mapping. Next, the features are reshaped to adapt to the subsequent fusion size, and the feature map data is normalized. After that, a traditional fusion feature map is generated, and the feature map based on hyperbolic space mapping is fused with the traditional feature map. Finally, the feature detection head decouples the fused feature map for classification and position regression prediction. After completing the loss calculation, the model training is completed.

[0111] In step S6, it also includes obtaining the simulated prediction results of the remote sensing image target detection model, and using the difference between the simulated prediction results and the real annotations of the input remote sensing image as a judgment basis to control the training.

[0112] During the training process of remote sensing image target detection models, human intervention can be used to timely discover and correct model errors, quickly identify error points and give correct suggestions and corrective measures. Human intervention can also be used to label and guide data. Through human assistance, more accurate labeling information can be provided for the model to improve the reliability and generation effect of the model.

[0113] Step S7: Detect the remote sensing image using the remote sensing image target detection model obtained in step S6.

[0114] In this embodiment, when training a remote sensing image target detection model based on a convolutional neural network, the input remote sensing image features are extracted through a feature extraction backbone network to obtain a hierarchical feature map, thereby improving the model's ability to classify and locate targets of interest in remote sensing images, and then a remote sensing image target detection method based on hyperbolic space mapping is used to project the shallow feature map, middle feature map, and deep feature map generated in the feature fusion network into the hyperbolic space to obtain geometric metric information of feature information in high-dimensional hyperbolic space, so as to help the target remote sensing image target detection model to utilize the spatial geometric metric information of high-dimensional feature data to assist prediction in the feature decoupling stage, thereby enhancing the information dimension of the original prediction feature map, helping the model to better understand the distribution of data, learn more abstract feature representations, and enhance the information dimension of the original prediction feature map, thereby improving the generalization ability of target detection tasks.

[0115] The remote sensing image target detection model can obtain multi-scale information of the target during the detection process, and obtain richer and more accurate target information, thereby improving the model's ability to classify and locate targets of interest in remote sensing images, and improving the accuracy of remote sensing image target positioning, which is of great significance for high-resolution remote sensing image rotating frame target detection.

[0116] Another embodiment of the present invention provides a remote sensing image target detection method based on hyperbolic space mapping, and its specific steps are as follows: Figure 4 As shown in the figure: step S100, obtain remote sensing image data, and perform appropriate preprocessing operations, including image cropping, image flipping, projection transformation, etc.; step S200, build a deep neural network model, and extract target features through the backbone feature extraction network; step S300, build a feature fusion network, and fuse the feature maps obtained by the backbone feature extraction network; step S400, use the hyperbolic space feature mapping algorithm to project the channel information of the multi-layer feature map obtained by the feature fusion network into the hyperbolic space to obtain the geometric metric information of the feature information in the high-dimensional hyperbolic space. Information; Step S500, the feature map obtained by the hyperbolic space mapping algorithm is spliced ​​with the unprojected normal feature map to use the geometric metric information of the hyperbolic space to assist the model in prediction; Step S600, the final feature map obtained after the feature detection head fusion is constructed for decoupling, and the classification loss and position prediction loss are calculated; Step S700, determine whether the training is completed, if so, execute step S800, if not, execute steps S100 to S600 again; Step S800, use the detection model obtained in step S700 to detect the remote sensing image.

[0117] According to one aspect of the present invention, there is provided an electronic device, comprising: one or more processors, one or more memories, and one or more computer programs; wherein the processor is connected to the memory, the one or more computer programs are stored in the memory, and when the electronic device is running, the processor executes the one or more computer programs stored in the memory, so that the electronic device performs the above-mentioned remote sensing image target detection method based on hyperbolic space mapping.

[0118] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements.

[0119] Finally, it should be noted that the above is a preferred embodiment of the present invention. It should be pointed out that although the preferred embodiment of the present invention has been described, for those skilled in the art, once the basic creative concept of the present invention is known, several improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications should also be regarded as the protection scope of the present invention. Therefore, the attached claims are intended to be interpreted as including the preferred embodiment and all changes and modifications that fall within the scope of the embodiments of the present invention.

Claims

1. A remote sensing image target detection method based on hyperbolic space mapping, characterized in that: The following steps are involved: Step S1, extracting features of an input remote sensing image through a feature extraction backbone network, and obtaining a plurality of hierarchical feature maps corresponding to the convolutional levels of the feature extraction backbone network; Step S2: using a feature fusion network, according to the type of the hierarchical feature map, the hierarchical feature maps of the same type are fused to obtain a first plurality of hierarchical feature maps of the remote sensing image target, wherein the types of the hierarchical feature maps include at least a deep feature map, a middle feature map, and a shallow feature map; Step S3, using a hyperbolic space mapping network to project the channel information of the plurality of the first multi-level feature maps into a hyperbolic space, to obtain a plurality of second multi-level feature maps, and to obtain geometric metric information of the feature information in a high-dimensional hyperbolic space; Step S4, concatenating the first multi-level feature map and the second multi-level feature map to obtain a plurality of fused feature maps; Step S5: construct a feature detection head, detect the fused feature map, and calculate the classification loss and the position prediction loss; Step S6, repeating steps S1 to S5 to train a remote sensing image target detection model, wherein the remote sensing image target detection model includes the feature extraction backbone network, the feature fusion network, the hyperbolic space mapping network and the feature detection head; Step S7: Detect the remote sensing image using the remote sensing image target detection model obtained in step S6.

2. The method for remote sensing image target detection based on hyperbolic space mapping according to claim 1, characterized in that: In the step S1, it specifically includes: Step S11, constructing the feature extraction backbone network; Step S12, performing a preprocessing operation on the input remote sensing image; the preprocessing operation includes image cropping, image flipping, projection transformation, mosaic enhancement and / or image filling; Step S13: performing feature extraction operation on the preprocessed remote sensing image through the feature extraction backbone network to obtain the hierarchical feature map.

3. The method for remote sensing image target detection based on hyperbolic space mapping according to claim 2, characterized in that: In the step S2, it specifically includes: Step S21, constructing the feature fusion network; Step S22: receiving the hierarchical feature map through the feature fusion network, and performing a fusion operation on the hierarchical feature maps of the same type according to the type of the hierarchical feature map to obtain the first multi-layer feature map; the fusion operation includes upsampling, lateral connection and / or dimensional splicing.

4. The method for remote sensing image target detection based on hyperbolic space mapping according to claim 3, characterized in that: In the step S22, specifically including: Step S221, traversing all the hierarchical feature maps, and performing channel number consistency processing on the hierarchical feature maps; Step S222: perform a size upsampling operation on the hierarchical feature map after the channel number consistency processing, adjust the size of the shallow feature map to the size of the middle feature map, and then perform dimensional splicing or channel addition to obtain the first multi-level feature map.

5. The method for remote sensing image target detection based on hyperbolic space mapping according to claim 4, characterized in that: In step S3, the hyperbolic space mapping network maps the first multi-level feature map vector by a hyperbolic space mapping algorithm, performs pixel-level data traversal on the channel information of the first multi-level feature map, extracts feature vectors of 1*1 channels according to pixel cycles, obtains a corresponding number of feature points, and then performs hyperbolic space mapping processing on the obtained feature points to obtain feature vectors; The hyperbolic space mapping algorithm is modeled as a Möbius transform or a Poincare disk.

6. The method for remote sensing image target detection based on hyperbolic space mapping according to claim 5, characterized in that: In step S3, the channel information of the feature map is projected into the hyperbolic space using the Mobius transform, which specifically includes: Step S31: Perform vector representation on the first multi-level feature map. For any first multi-level feature map, define F∈R H×W×D is the first multi-level feature map from the feature fusion network, with a size of H×W and a number of channels of D; F=[x1,x2,…,x i ,…,x N ] Among them, [x1,x2,…,x i ,…,x N ] represents the sequence data obtained after the first multi-level feature map F is stretched, N = H × W, x i ∈R N*D Represents the two-dimensional vector corresponding to the feature point at the i-th position in the first multi-level feature map, i∈[1,2,…,N]; Step S32: Map the channel information of the plurality of the first multi-level feature maps in the Euclidean distance space to the Riemann sphere space using the Mobius transformation method to obtain data structure property information in the surface space. The overall overview of the Mobius transformation method is as follows: Define the extended complex plane The Mobius transform of , where С is the complex plane, ∞ is the point at infinity, for the two-dimensional vector x corresponding to the feature point at the i-th position in the first multi-level feature map F i =[N i ,D i ], after the Möbius transformation τ(·), we get: Where S is a reversible matrix used to represent the projection transformation parameters a, b, c, d: For the two-dimensional vector x corresponding to the feature point at the i-th position in the first multi-level feature map F i , the universal Möbius transformation τ(x i ) is expressed as follows: The Möbius transform remains unchanged by the scaling of the projection transformation parameters, and its fixed point γ can be obtained by solving: After the Möbius transformation, the feature matrix F of the similar shape of the first multi-level feature map F is obtained. Mobius , its internal data is characterized by the feature mapping of the feature data of the first multi-level feature map F in the Riemann sphere space, that is: F Mobius =[x Mobius,1 ,x Mobius,2 ,…,x Mobius,i ,…,x Mobius,M ] where F Mobius is the feature matrix of the second multi-level feature map obtained by transforming the first multi-level feature map F, [x Mobius,i ] means [x i ] corresponds to the transformation structure, M represents the sequence number of the feature data obtained by the transformation.

7. The method for remote sensing image target detection based on hyperbolic space mapping according to claim 6, characterized in that: The step S4 specifically includes: Step S41, reshaping the plurality of second multi-level feature maps obtained in step S3; Step S42, performing feature normalization on the plurality of second multi-level feature maps obtained after feature reshaping in step S41, and normalizing them to a range of 0 to 255; Step S43, using the normalized second multi-level feature maps to fuse with the first multi-level feature maps in step S2 to obtain a plurality of fused feature maps; wherein the fusion method includes an addition operation, a feature concatenation operation and / or a dual fusion operation; Step S44, filtering the multiple fused feature maps obtained in step S43 through an attention mechanism to obtain an optimal channel information representation; Step S45, using the fused feature map screened out in step S44 to transmit and send to the feature detection head.

8. The method for remote sensing image target detection based on hyperbolic space mapping according to claim 7, characterized in that: In the step S41, it specifically includes: The feature matrix F of the second multi-level feature map obtained after step S33 is reshaped by the reshape operation Mobius Process it and normalize it to the first multi-level feature map F∈R H×W×D within the domain.

9. The method for remote sensing image target detection based on hyperbolic space mapping according to claim 1, characterized in that: In step S6, the simulation prediction result of the remote sensing image target detection model is obtained, and the difference between the simulation prediction result and the real annotation of the input remote sensing image is used as a judgment basis to control the training.

10. An electronic device, characterized in that: include: One or more processors, one or more memories, and one or more computer programs; wherein the processor is connected to the memory, and the one or more computer programs are stored in the memory. When the electronic device is running, the processor executes the one or more computer programs stored in the memory, so that the electronic device performs the remote sensing image target detection method based on hyperbolic space mapping as described in any one of claims 1 to 9.

Images