CN-121995366-A - Unmanned aerial vehicle high-precision interception method and system based on target detection algorithm

CN121995366ACN 121995366 ACN121995366 ACN 121995366ACN-121995366-A

Abstract

The invention provides a high-precision interception method and a high-precision interception system for an unmanned aerial vehicle based on a target detection algorithm, which relate to the technical field of unmanned aerial vehicle identification and comprise the steps of collecting flight image data and radar echo data of a target unmanned aerial vehicle, and preprocessing to obtain standardized image data and standardized radar data; the method comprises the steps of constructing a target detection model based on a target detection algorithm, converting an initial positioning result of a target unmanned aerial vehicle under an image coordinate system into initial position data of the target unmanned aerial vehicle under a geographic coordinate system according to the target detection model, standardized image data and standardized radar data, carrying out state correction on the initial position data of the target unmanned aerial vehicle based on a Hungary algorithm to obtain corresponding high-precision position data, carrying out interception path analysis on the high-precision position data and radar echo data based on an improved A-type algorithm to obtain an optimal interception path of the target unmanned aerial vehicle, and further executing corresponding interception execution instructions.

Inventors

WEI QINGSONG
LIU XI
WANG MINGLING
YUAN ZISHUN
Zhu Lvtao

Assignees

浙江理工大学

Dates

Publication Date: 20260508
Application Date: 20260119

Claims (8)

1. The unmanned aerial vehicle high-precision interception method based on the target detection algorithm is characterized by comprising the following steps of: The method comprises the steps of collecting flight image data and radar echo data of a target unmanned aerial vehicle in a monitoring area, preprocessing the flight image data and the radar echo data, and obtaining corresponding standardized image data and standardized radar data; according to the target detection model, the standardized image data and the standardized radar data, converting an initial positioning result of the target unmanned aerial vehicle under an image coordinate system into initial position data of the target unmanned aerial vehicle under a geographic coordinate system; based on a Hungary algorithm, performing state correction on initial position data of the target unmanned aerial vehicle to obtain high-precision position data of the target unmanned aerial vehicle; and based on an improved A-algorithm, performing interception path analysis on the high-precision position data and the radar echo data of the target unmanned aerial vehicle to obtain an optimal interception path of the target unmanned aerial vehicle, and further executing a corresponding interception execution instruction.
2. The unmanned aerial vehicle high-precision interception method based on the target detection algorithm according to claim 1, wherein the process of constructing the target detection model based on the target detection algorithm comprises the following steps: Acquiring a plurality of groups of standardized image data with different models, different flight attitudes and aligned unmanned aerial vehicle time stamps under different environments and standardized radar data to form a data set; Based on YOLOv8 as a basic algorithm frame, improving the YOLOv basic algorithm frame to obtain an improved YOLOv algorithm frame; Based on the training set, performing three-stage training on the improved YOLOv algorithm frame to obtain a trained improved YOLOv algorithm frame; And checking the trained improved YOLOv algorithm framework, and obtaining a corresponding target detection model after checking.
3. The method for high-precision interception of unmanned aerial vehicle based on object detection algorithm according to claim 2, wherein the process for improving YOLOv basic algorithm framework comprises: Adding a light coordinate attention mechanism module to a backbone feature extraction network in a YOLOv basic algorithm framework, and adjusting the number of feature channels of the backbone feature extraction network; Hierarchical optimization is carried out on the neck feature fusion network in the YOLOv basic algorithm framework, wherein the hierarchical optimization comprises a pyramid feature fusion structure, a middle layer and a feature fusion residual connection layer, and the number of feature channels of the neck feature fusion network is adjusted.
4. The method for high-precision interception of unmanned aerial vehicle based on target detection algorithm according to claim 3, wherein the process of performing three-stage training on the improved YOLOv algorithm frame based on training set comprises: Loading a data set, performing data enhancement configuration, setting super-parameter configuration, and completing training iteration for corresponding times or detecting indexes corresponding to the training set in the data set to be consistent, and performing second-stage training; The second stage training comprises the steps of improving the calling proportion of unmanned aerial vehicle small target samples, night infrared samples, haze scene samples and fuselage part shielding samples in a data set, dynamically regulating and controlling super parameters, optimizing weight preservation frequency, completing training iteration of corresponding times or meeting detection indexes corresponding to a training set and a testing set in the data set, and performing third stage training; The third-stage training comprises the steps of loading unmanned aerial vehicle targets in a data set, enabling the unmanned aerial vehicle targets to be free of shielding, clear in outline and complete in characteristics, enabling a scene to be a calling duty ratio of a single illumination environment, completely closing all data enhancement configurations, enabling learning rate to execute secondary step-like attenuation, keeping all optimizer parameters unchanged, completing training iteration for corresponding times, enabling detection indexes corresponding to a training set and a testing set in the data set to be consistent, or triggering an early-stopping mechanism, and completing the third-stage training.
5. The method for high-precision interception of a target unmanned aerial vehicle based on a target detection algorithm according to claim 4, wherein the process of converting the initial positioning result of the target unmanned aerial vehicle in an image coordinate system into the initial position data of the target unmanned aerial vehicle in a geographic coordinate system according to the target detection model, the standardized image data and the standardized radar data comprises the following steps: Inputting standardized image data and standardized radar data into a trained target detection model, outputting boundary frame coordinates of the target unmanned aerial vehicle under an image coordinate system, calculating boundary frame center point coordinates, and taking the boundary frame center point coordinates as an initial positioning result under the image coordinate system; and converting the coordinates of the center point of the boundary frame under the image coordinate system into initial position data under the geographic coordinate system through a coordinate conversion formula according to the target distance, the azimuth angle and the installation parameters of the camera in the corresponding standardized radar data.
6. The unmanned aerial vehicle high-precision interception method based on the target detection algorithm according to claim 5, wherein the process of performing state correction on the initial position data of the target unmanned aerial vehicle based on the hungarian algorithm comprises the following steps: If the state correction is needed, constructing a cost matrix based on a Hungary algorithm, further obtaining corresponding matching cost, setting a space distance threshold, determining the optimal association relation of the front frame position data and the rear frame position data according to the space distance threshold and the corresponding matching cost, further carrying out smooth correction on the initial position data of the optimal association relation according to the corresponding radial speed, and obtaining corresponding high-precision position data.
7. The unmanned aerial vehicle high-precision interception method based on the target detection algorithm according to claim 6, wherein the process of performing interception path analysis on the high-precision position data and the radar echo data of the target unmanned aerial vehicle based on the improved a-x algorithm comprises the following steps: Constructing a three-dimensional grid map by taking geographic coordinates of a monitoring area as a reference, setting the grid size of the three-dimensional grid map, marking the no-fly area as an obstacle grid, and setting constraint conditions; The method comprises the steps of introducing radial speed in radar echo data into a heuristic function of an algorithm A to obtain an improved algorithm A, creating two node sets based on the improved algorithm A, carrying out serial search according to priority by taking current high-precision position data of the interception unmanned aerial vehicle as a starting point, carrying out constraint verification according to constraint conditions to generate an optimal interception path of the target unmanned aerial vehicle, if the constraint verification is satisfied, the corresponding path is the optimal interception path, if the constraint verification is not satisfied, adjusting parameters to reprogram, dividing the optimal interception path according to the flight state of the target unmanned aerial vehicle, and executing corresponding interception execution instructions.
8. The unmanned aerial vehicle high-precision interception system based on the target detection algorithm realizes the unmanned aerial vehicle high-precision interception method based on the target detection algorithm according to any one of claims 1 to 7, and is characterized by comprising a data acquisition module, a data preprocessing module, a target detection module, a target correction module and an intelligent management module; the data acquisition module is used for acquiring flight image data and radar echo data of the target unmanned aerial vehicle in the monitoring area; the data preprocessing module is used for preprocessing the flight image data and the radar echo data to obtain corresponding standardized image data and standardized radar data; the target detection module is used for constructing a target detection model based on a target detection algorithm, and converting an initial positioning result of the target unmanned aerial vehicle under the image coordinate system into initial position data of the target unmanned aerial vehicle under the geographic coordinate system according to the target detection model, the standardized image data and the standardized radar data; the target correction module is used for carrying out state correction on initial position data of the target unmanned aerial vehicle based on the Hungary algorithm to obtain high-precision position data of the target unmanned aerial vehicle; And the intelligent management module is used for carrying out interception path analysis on the high-precision position data and the radar echo data of the target unmanned aerial vehicle based on an improved A-based algorithm to obtain an optimal interception path of the target unmanned aerial vehicle, and further executing a corresponding interception execution instruction.

Description

Unmanned aerial vehicle high-precision interception method and system based on target detection algorithm Technical Field The invention relates to the technical field of unmanned aerial vehicle identification, in particular to an unmanned aerial vehicle high-precision interception method and system based on a target detection algorithm. Background The unmanned aerial vehicle interception task needs to realize accurate positioning and efficient interception of the target unmanned aerial vehicle, an effective state correction mechanism is lacked after the target is positioned, the position error can directly influence the accuracy of the subsequent interception path planning, only distance factors are considered, the motion characteristics of the target unmanned aerial vehicle are not fully combined, the planned path possibly has the problems of infeasibility in dynamics, inaccurate interception opportunity control and the like, and the actual requirement of high-precision interception is difficult to meet. Therefore, the unmanned aerial vehicle high-precision interception method and system based on the target detection algorithm are provided. Disclosure of Invention In order to solve the technical problems, the invention aims to provide an unmanned aerial vehicle high-precision interception method and system based on a target detection algorithm. In order to achieve the purpose, the invention provides the following technical scheme that the unmanned aerial vehicle high-precision interception method based on the target detection algorithm comprises the following steps: The method comprises the steps of collecting flight image data and radar echo data of a target unmanned aerial vehicle in a monitoring area, preprocessing the flight image data and the radar echo data, and obtaining corresponding standardized image data and standardized radar data; according to the target detection model, the standardized image data and the standardized radar data, converting an initial positioning result of the target unmanned aerial vehicle under an image coordinate system into initial position data of the target unmanned aerial vehicle under a geographic coordinate system; based on a Hungary algorithm, performing state correction on initial position data of the target unmanned aerial vehicle to obtain high-precision position data of the target unmanned aerial vehicle; and based on an improved A-algorithm, performing interception path analysis on the high-precision position data and the radar echo data of the target unmanned aerial vehicle to obtain an optimal interception path of the target unmanned aerial vehicle, and further executing a corresponding interception execution instruction. Further, based on the target detection algorithm, the process of constructing the target detection model includes: Acquiring a plurality of groups of standardized image data with different models, different flight attitudes and aligned unmanned aerial vehicle time stamps under different environments and standardized radar data to form a data set; Based on YOLOv8 as a basic algorithm frame, improving the YOLOv basic algorithm frame to obtain an improved YOLOv algorithm frame; Based on the training set, performing three-stage training on the improved YOLOv algorithm frame to obtain a trained improved YOLOv algorithm frame; And checking the trained improved YOLOv algorithm framework, and obtaining a corresponding target detection model after checking. Further, the process of improving YOLOv basic algorithm framework includes: Adding a light coordinate attention mechanism module to a backbone feature extraction network in a YOLOv basic algorithm framework, and adjusting the number of feature channels of the backbone feature extraction network; Hierarchical optimization is carried out on the neck feature fusion network in the YOLOv basic algorithm framework, wherein the hierarchical optimization comprises a pyramid feature fusion structure, a middle layer and a feature fusion residual connection layer, and the number of feature channels of the neck feature fusion network is adjusted. Further, based on the training set, the process of performing three-stage training on the improved YOLOv algorithm framework comprises the following steps: Loading a data set, performing data enhancement configuration, setting super-parameter configuration, and completing training iteration for corresponding times or detecting indexes corresponding to the training set in the data set to be consistent, and performing second-stage training; The second stage training comprises the steps of improving the calling proportion of unmanned aerial vehicle small target samples, night infrared samples, haze scene samples and fuselage part shielding samples in a data set, dynamically regulating and controlling super parameters, optimizing weight preservation frequency, completing training iteration of corresponding times or meeting detection indexes corresponding to a training se