摘要: |
无人机巡线作为检测电力线的重要手段,在飞行过程中准确识别障碍物是保障巡线任务可靠完成的关键。但目前对无人机巡线过程中常见障碍物,如电力线、电力塔、树木的识别受恶劣天气环境干扰严重,致使误判、漏判。为此,基于毫米波雷达传感器具有不受天气、光线因素影响,复杂环境中工作稳定等特点,本文提出基于毫米波雷达的无人机障碍物分类方法。该方法首先通过毫米波雷达采集三类障碍物的原始数据并提取其距离-速度Doppler及距离-方位角Doppler信息,接着分别通过特征值分解及共生灰度矩阵实现特征提取,最后通过蛇鹭优化算法实现对三类障碍物的目标分类。实验结果表明,本文方法对电力线、电力塔和树木的整体识别准确率达89.4%,与传统方法相比具有较高的识别准确率及鲁棒性。 |
关键词: 毫米波雷达 障碍物分类 特征提取 蛇鹭优化算法 |
DOI: |
分类号:TM755;TP212.9 |
基金项目:本项目受湖北省自然科学基金 (2022CFA007) 、武汉市知识创新专项(No.2022020801010258)项目资助 |
|
Millimeter-wave Radar-based Obstacle Classification Method for Unmanned Aerial Vehicles |
|
|
Abstract: |
As an important means of detecting power lines, accurate identification of obstacles during flight is the key to guaranteeing the reliable completion of the patrol task. However, at present, the identification of common obstacles during UAV patrol, such as power lines, power towers and trees, is seriously interfered by the bad weather environment, resulting in misjudgment and omission of judgment. For this reason, based on the millimeter-wave radar sensor has the characteristics of not affected by weather and light factors, and stable work in complex environments, this paper proposes a UAV obstacle classification method based on millimeter-wave radar. The method firstly collects the raw data of three types of obstacles and extracts their distance-velocity Doppler and distance-azimuth Doppler information through millimeter wave radar, then realizes feature extraction through eigenvalue decomposition (SVD) and symbiotic gray matrix (GLCM) respectively, and finally realizes the target classification of the three types of obstacles through the snake and heron optimization algorithm (SBOA). The experimental results show that the overall recognition accuracy of this paper"s method for power lines, power towers and trees reaches 89.4%, which has high recognition accuracy and robustness compared with traditional methods. |
Key words: millimeter wave radar obstacle classification feature extraction SBOA |