Unmanned aerial vehicle echo characteristic modeling method based on clustering algorithm

Abstract

The invention relates to the field of data processing, in particular to an unmanned aerial vehicle echo characteristic modeling method based on a clustering algorithm, and the method comprises the steps: extracting characteristic parameters from an unmanned aerial vehicle radar echo original signal, and mapping the characteristic parameters into two-dimensional data points; carrying out region block division and communication block generation on the two-dimensional data points; when the number of the connected blocks meets a first preset condition, the centroid of each connected block is calculated, and the data point closest to the centroid serves as the initial center point of the connected block; when the number of the communication blocks meets a second preset condition, selecting the communication blocks for splitting until the number of all the communication blocks meets a first preset condition, and finally determining an initial center point of the communication blocks; and clustering the two-dimensional data points by using the initial center point, and establishing a radar cross section model according to the center point of each cluster. The number of the communication blocks and selection of the initial center points are dynamically adjusted, it is ensured that selection of the initial center points is more reasonable, and the clustering effect is improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing. More specifically, the present invention relates to a method for modeling the echo characteristics of an unmanned aerial vehicle (UAV) based on a clustering algorithm. Background Art

[0002] Currently, the application of UAVs in fields such as geographical mapping and film aerial photography has been quite mature, and they are also playing an increasingly crucial role in emerging scenarios such as emergency rescue. However, with the rapid popularization of UAVs, the phenomena of "illegal flight" and "unregulated flight" occur frequently, posing a serious threat to public safety.

[0003] Traditional UAV detection mainly relies on the combination of radar detection guidance and optoelectronic confirmation. Although radar has the advantages of long detection range and high efficiency, it is easily affected by factors such as ground clutter, multipath effects, and birds, resulting in an increase in the false alarm rate.

[0004] UAV echo data has extremely high complexity and contains a large number of sampling points. The K-means clustering algorithm can be used to simplify these complex data into several clusters, making the data structure clearer and facilitating subsequent analysis and modeling work. However, in the traditional K-means clustering algorithm, the initial center points are randomly selected. In terms of the echo characteristics of UAVs, if both initial center points are located in the low-intensity echo area and are close to each other, some medium-intensity echo data may be wrongly classified into these two clusters and cannot form independent clusters. If the initial center points are selected near the edge or outliers of the data distribution, deformed clustering results will also be generated. As the input data for modeling, incorrect clustering leads to insufficient ability of the model to distinguish normal and abnormal echoes, which affects the subsequent analysis of the flight state of UAVs. Summary of the Invention

[0005] To solve the technical problem that the traditional K-means clustering algorithm is prone to incorrect clustering when processing UAV echo data due to randomly selected initial center points, which affects subsequent modeling, the present invention provides the following technical solutions.

[0006] A method for modeling the echo characteristics of an unmanned aerial vehicle based on a clustering algorithm, comprising: extracting characteristic parameters from the original radar echo signal of the UAV and mapping the characteristic parameters into two-dimensional data points; performing regional block division and connected component generation on the two-dimensional data points; When the number of connected components satisfies the first preset condition, calculate the centroid of each connected component, and use the data point closest to the centroid as the initial center point corresponding to the connected component; when the number of connected components satisfies the second preset condition, select a connected component for splitting until the number of all connected components satisfies the first preset condition, then stop splitting, calculate the centroid of each connected component, and use the data point closest to the centroid as the initial center point corresponding to the connected component. Perform clustering operations on the two-dimensional data points using the initial center points to obtain multiple clustering clusters, and establish a radar cross-section model based on the center points of each clustering cluster.

[0007] Preferably, the characteristic parameters are signal amplitude and phase.

[0008] Preferably, the generation of the connected components includes: For each region block, count the number of data points falling into the region block, and perform binary classification on all region blocks. Mark the category of region blocks with the largest average number of points as 1, and mark the other category of region blocks as 0. Merge the region blocks marked as 1 into one connected component according to the spatial neighborhood relationship, and thus obtain all independent connected components.

[0009] Preferably, calculate the areas of all connected components, and select the connected component with the largest area for splitting.

[0010] Preferably, if there are multiple connected components with the same and largest area values, calculate the priorities corresponding to these connected components with the largest area, and select the connected component with the largest priority for splitting.

[0011] Preferably, the splitting process includes: Map the selected connected component to be split into a point set, that is, use the centroid of each region block included in the connected component as a point, and perform binary classification on the point set. Select the two points with the farthest distance in the point set as the initial center points for binary classification. After the binary classification is completed, calculate the difference in the number of points between the two categories. If the difference in the number of points is greater than or equal to a preset value, for each point in the category with more points, calculate its distance to the center point of its own category, and calculate the distance from the points in its own category to the center point of the other category; for each point, calculate the difference between its distance to the center point of its own category and its distance to the center point of the other category, sort the calculated differences from largest to smallest, and reassign the points ranked last and equal to the number of the difference in the number of points to the other category to complete the splitting of the connected component.

[0012] Preferably, the process of obtaining the priority includes: Calculate the aspect ratio, filling degree, and corner connection degree of each connected component among all connected components with the same and largest area values, and use the product of the aspect ratio, filling degree, and corner connection degree as the priority.

[0013] Preferably, it further includes: When the number of connected components meets the third preset condition, use a clustering algorithm to cluster the centroids of the connected components to obtain k virtual connected components, and then obtain k initial center points.

[0014] The beneficial effects of the present invention are: The present invention first retains the key information of the echo signal through dimensionality reduction, focuses on the core features of the signal, and provides intuitive input for clustering; then filters noise and sparse regions through density segmentation, initially locates potential clustering centers, avoids the deviation caused by noise or outliers in traditional clustering algorithms, and further performs different operations according to whether the number of connected components meets the preset conditions, dynamically adjusts the number of connected components and the selection of initial center points to meet the requirements of subsequent clustering operations, ensures the more reasonable selection of initial center points, and improves the clustering effect.

[0015] Establish a radar cross-section model based on the center point of each clustering cluster. The center point of the clustering cluster represents the average characteristics of the data points within the cluster, and the radar cross-section model established based on these center points can reflect different modes of the UAV radar echo characteristics, realizing effective modeling of the UAV echo characteristics. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a flowchart of the method of steps S1 - S4 in a method for modeling UAV echo characteristics based on a clustering algorithm according to an embodiment of the present invention.

[0017] Figure 2 is a schematic diagram of the marked result of the regional block in a method for modeling UAV echo characteristics based on a clustering algorithm according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present application.

[0019] Refer to Figure 1 , a method for modeling UAV echo characteristics based on a clustering algorithm includes steps S1 - S4, specifically as follows: S1: Extract feature parameters from the original radar echo signal of the UAV and map the feature parameters into two-dimensional data points.

[0020] In one embodiment, a radar device is used to detect the drone, receive the original reflected signal (i.e., the echo data of the electromagnetic wave), and further extract key features such as signal amplitude, phase, frequency, and delay from the original signal. The extracted feature parameters are divided into two groups, each with two parameters, to form two two-dimensional coordinate spaces.

[0021] For example, space 1: the horizontal axis is signal amplitude, and the vertical axis is phase, which describes the reflection characteristics of the target; space 2: the horizontal axis is frequency, and the vertical axis is delay, which describes the motion state and distance of the target.

[0022] For example, the amplitude extracted from the original radar echo signal of a drone in a certain measurement is , the phase is , then draw the point in space 1 The frequency is , the delay is , then draw the point in space 2 .

[0023] Through the above processing, the originally complex radar echo raw signal can be converted into clusterable two-dimensional data points, which is convenient for analyzing and processing the UAV radar echo model.

[0024] S2: Divide the two-dimensional data points into regional blocks and generate connected blocks.

[0025] To further analyze the distribution of data points in S1, take space 1 in S1 as an example. In one embodiment, assuming that data points need to be classified into k categories, the entire space is divided into k×k small square blocks. The size of each block is determined by the data range (coordinate range) and the value of k.

[0026] Then, for each area block, the number of data points falling into the area block is counted, and the K-means clustering algorithm is used to divide all area blocks into two categories according to the number of their data points, one category represents high-density area blocks (i.e., more points), and the other category represents low-density area blocks (i.e., fewer points). Then, the clustering result is obtained, that is, each area block is assigned to a category.

[0027] Next, for the above clustering results, the area block with the largest average number of points is marked as 1, and the other area blocks are marked as 0. The area blocks marked as 1 are merged into larger connected areas according to the spatial neighborhood relationship. The marking results are as follows: Figure 2 shown.

[0028] Exemplarily, each area block marked as 1 is checked. If its four neighbors (four adjacent blocks above, below, left, and right) are also 1, these blocks are merged into a connected block, and finally multiple independent connected blocks are obtained.

[0029] The original data may contain a large number of discrete points, and directly performing clustering on these points involves a huge computational amount. After the above operations, by dividing the space into regional blocks and classifying and merging the regional blocks, the number of data units that need to be processed can be significantly reduced, thereby reducing the computational complexity.

[0030] S3: When the number of connected components meets the first preset condition, calculate the centroid of each connected component, and use the data point closest to the centroid as the initial center point corresponding to the connected component; when the number of connected components meets the second preset condition, select a connected component for splitting until the number of all connected components meets the first preset condition, then stop splitting, calculate the centroid of each connected component, and use the data point closest to the centroid as the initial center point corresponding to the connected component.

[0031] The traditional K-means clustering algorithm needs to pre-specify the number of clustering categories k and depends on the position of the initial center points. The selection of the initial center points directly affects the quality of the clustering results and the convergence speed of the algorithm.

[0032] If the number of initial center points (i.e., the number of connected components) is less than k, the K-means clustering cannot be started. Conversely, if the number of initial center points (i.e., the number of connected components) is greater than k, additional processing is required to select k representative center points.

[0033] In one embodiment, different processing methods are formulated based on the number of connected components. When the number of connected components meets the first preset condition (i.e., the number of connected components is equal to k), calculate the centroid of each connected component, and use the data point closest to the centroid as the initial center point corresponding to the connected component.

[0034] When the number of connected components meets the second preset condition (i.e., the number of connected components is less than k), directly selecting the centroid of the connected component will result in insufficient initial center points. Therefore, new initial center points can be generated by splitting the connected component to increase diversity.

[0035] Exemplarily, calculate the areas of all connected components, select the connected component with the largest area for splitting, and the splitting operation is as follows: Map the connected component selected for splitting to a point set, that is, the centroid of each regional block included in the connected component is used as a point, and the point set is binary-classified, where the two points with the farthest distance in the point set are selected as the initial center points for binary classification.

[0036] After the binary classification is completed, calculate the difference in the number of points between the two categories, denoted as d. If d is greater than or equal to 2, for each point in the category with more points, calculate its distance to the center point of its own category. At the same time, calculate the distance from this point to the center point of the other category. Then, for each point, calculate the difference between the distance to the center point of its own category and the distance to the center point of the other category. Sort the calculated differences from largest to smallest, and reassign the last d points to the other category to balance the number of points in the two categories. Finally, split the original connected component into two new connected components (each category corresponds to a new connected component).

[0037] In addition, if d is less than 2, it means that the result of the binary classification is already balanced enough and no further adjustment is needed. Just divide the region block corresponding to the points into two new connected components directly according to the result.

[0038] After the splitting is completed, re - count the number of connected components. If the number of connected components still meets the above - mentioned second preset condition, repeat the above splitting operation until the number of connected components meets the above - mentioned first preset condition.

[0039] It should be noted that if there are multiple connected components with the same and largest area value, calculate the priorities corresponding to these connected components with the largest area, and select the connected component with the largest priority for splitting.

[0040] The calculation formula for the above - mentioned priority is as follows:

[0041] In the formula, is the priority of the connected component, and are the length and width of the circumscribed rectangle of the connected component respectively, is the area of the connected component, is the number of corner points shared by the connected component and other connected components, is the number of corner points on the outermost boundary of the connected component.

[0042] Among them, represents the aspect ratio of the connected component. In clustering, slender connected components are more difficult to process. Therefore, it is also necessary to adjust the contribution direction of the aspect ratio, that is, use . Then, when is smaller, is larger, and the priority is larger, indicating that slender connected components are split more preferentially; represents the filling degree of the connected component, that is, the compactness of the connected component in the circumscribed rectangle. Compact connected components (i.e., is large) usually do not need to be split preferentially because their shapes are more regular. Similarly, use . When is smaller, The larger it is, the higher the priority The larger it is, indicating that the connected components with sparse filling are preferentially split; Indicates the degree of connection of the connected component with other connected components at the corner points, indicating that the connected component has more connections with other connected components at the corner points, The larger it is, the higher the priority The larger it is, the more it needs to be split preferentially.

[0043] It should be noted that when the number of connected components meets the third preset condition (that is, the number of connected components is greater than k), in this case, the centroid coordinates of each connected component are used as data points, and the traditional K-means clustering algorithm is used to cluster these centroid points into k categories. Each clustering result corresponds to a "virtual connected component", and its center is the centroid point of this category, and finally k connected components are obtained, meeting the requirements of the initial center point quantity.

[0044] According to the above operations, all connected components are finally obtained, and the number of connected components is equal to k, and the centroid of each connected component is calculated, and the data point closest to the centroid is used as the initial center point corresponding to the connected component.

[0045] According to the operations of S2 and S3 above, all initial center points in space 2 can be obtained in the same way.

[0046] S4: Use the initial center points to perform clustering operations on the two-dimensional data points to obtain multiple clustering clusters, and establish a radar cross-section model according to the center points of each clustering cluster.

[0047] According to all the initial center points obtained above, these initial center points are used as the initial center points for the K-means clustering algorithm operation, and k clusters are obtained by clustering. Calculate the center point of each cluster, which represents the typical characteristics of the cluster, and select a suitable mathematical model (such as polynomial, exponential function, etc.) according to the characteristics of the cluster center to fit the relationship between the radar cross-section (RCS) and the input parameters, and establish a corresponding radar cross-section model for each clustering cluster to predict the echo characteristics of different types of unmanned aerial vehicles under different observation angles and radar parameters.

[0048] In summary, the present invention first performs regional block division and connected component generation on two-dimensional data points, and then dynamically adjusts the selection of the initial center points according to the number of connected components. When the number of connected components meets the first preset condition, calculate the centroid of each connected component, and use the data point closest to the centroid as the initial center point; when the number of connected components meets the second preset condition, select the connected components for splitting until the first preset condition is met, and finally determine the initial center points. This method avoids the blindness of randomly selecting the initial center points in the traditional K-means algorithm and makes the selection of the initial center points more reasonable.

[0049] During the connected component splitting process, features such as the area, aspect ratio, filling degree, and corner connection degree of the connected components are considered, and the connected component with the largest area or the highest priority is preferentially selected for splitting. This splitting method based on the data distribution characteristics can more reasonably adjust the distribution of the connected components, ensure that the initial center points can cover the main distribution area of the data, and avoid generating abnormal clustering results.

[0050] By dynamically adjusting the number of connected components and the selection of the initial center points, the present invention improves the clustering effect. The reasonable selection of the initial center points makes the clustering results more accurate and can better reflect the internal laws and patterns of the UAV echo characteristics.

[0051] The technical features of the above-described embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0052] The above-described embodiments only represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the patent application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several deformations and improvements can be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the appended claims.

Claims

1. A method for modeling UAV echo characteristics based on clustering algorithm, characterized in that: include: Extract characteristic parameters from the original radar echo signal of the UAV and map the characteristic parameters into two-dimensional data points; Divide two-dimensional data points into regional blocks and generate connected blocks; When the number of connected blocks meets the first preset condition, the centroid of each connected block is calculated, and the data point closest to the centroid is used as the initial center point corresponding to the connected block; when the number of connected blocks meets the second preset condition, the connected blocks are selected for splitting until the number of all connected blocks meets the first preset condition, then the splitting is stopped, and the centroid of each connected block is calculated, and the data point closest to the centroid is used as the initial center point corresponding to the connected block; The two-dimensional data points are clustered using the initial center point to obtain multiple clusters, and a radar cross-section model is established according to the center point of each cluster.

2. The method for modeling UAV echo characteristics based on clustering algorithm according to claim 1, characterized in that: The characteristic parameters are signal amplitude and phase.

3. The method for modeling UAV echo characteristics based on clustering algorithm according to claim 2 is characterized in that: The connected block generation includes: For each area block, count the number of data points that fall into the area block, and classify all area blocks into two categories. Mark the area block with the largest average number of points as 1, and the other area blocks as 0. The area blocks marked as 1 are merged into a connected block according to the spatial neighborhood relationship, and then all independent connected blocks are obtained.

4. The method for modeling UAV echo characteristics based on clustering algorithm according to claim 3 is characterized in that: Calculate the areas of all connected blocks and select the connected block with the largest area for splitting.

5. The method for modeling UAV echo characteristics based on clustering algorithm according to claim 4 is characterized in that: If there are multiple connected blocks with the same area value and the largest area, calculate the priorities corresponding to these connected blocks with the largest area, and select the connected block with the largest priority for splitting.

6. The method for modeling UAV echo characteristics based on clustering algorithm according to claim 5, characterized in that: The splitting process includes: The connected block selected for splitting is mapped to a point set, that is, the centroid of each regional block contained in the connected block is taken as a point, and the point set is binary classified, wherein the two points with the farthest distance in the point set are selected as the initial center points of the binary classification; after the binary classification is completed, the point difference between the two categories is calculated. If the point difference is greater than or equal to the preset value, for each point in the category with more points, its distance to the center point of this category is calculated, and the distance from the point in this category to the center point of another category is calculated; for each point, the difference between its distance to the center point of this category and the distance to the center point of another category is calculated, and the calculated differences are sorted from large to small, and the last point with the same number of point difference is reallocated to another category to complete the splitting of the connected block.

7. The method for modeling UAV echo characteristics based on clustering algorithm according to claim 6, characterized in that: The priority acquisition process includes: The aspect ratio, fullness and corner point connectivity of each connected block among all connected blocks with consistent area values ​​and the largest value are calculated, and the product of the aspect ratio, fullness and corner point connectivity is used as the priority.

8. The method for modeling UAV echo characteristics based on clustering algorithm according to claim 7, characterized in that: Also includes: When the number of connected blocks meets the third preset condition, a clustering algorithm is used to cluster the centroids of the connected blocks to obtain k virtual connected blocks, and then to obtain k initial center points.

Images