• 首页
  • 期刊简介
  • 编委会
  • 版权声明
  • 投稿指南
  • 期刊订阅
  • 相关下载
    雷达数据
    下载专区
  • 过刊浏览
  • 联系我们
引用本文:[点击复制]
[点击复制]
【打印本页】   【下载PDF全文】   【查看/发表评论】  【下载PDF阅读器】  【关闭】
←前一篇|后一篇→ 过刊浏览    高级检索
本文已被:浏览 79次   下载 0次  
分享到: 微信 更多
字体:加大+|默认|缩小-
基于TX2的机载毫米波雷达高压线检测技术实现
周砚龙, 何晨阳, 厉梦雪
中国航空工业集团公司雷华电子技术研究所
摘要:
高压线因其体积小,难发现的特性,目前已成为低空飞行时的首要威胁。为应对低空复杂场景中传统高压线检测算法性能下降的问题,研究了一种基于TX2的高压线智能检测技术。该技术以卷积神经网络(CNN)为核心,构建了深度梯度网络模型(DGNET)和分组池化卷积神经网络模型(GPCNN)分别用于图像分割和峰值检测。以雷达回波数据为输入进行模型推理,通过后处理策略融合推理结果并提取高压线。基于TX2硬件平台和不同的任务调度机制完成了该高压线智能检测技术在机载毫米波雷达上的工程应用,试验结果表明该技术下的高压线检测性能稳健,有很好的实时性,能够满足工程应用的需求。
关键词:  高压线检测  深度学习  毫米波雷达  任务调度
DOI:
分类号:TN957
基金项目:
Implementation of High Voltage Line Detection Technology for Airborne Millimeter Wave Radar Based on TX2
Abstract:
High voltage lines, due to their small size and difficulty to detect, have become the primary threat to low-altitude flight in recent years. Due to performance decline of traditional high voltage line detection algorithms in complex low-altitude scenarios, this paper has developed an intelligent detection technology based on TX2 to address the problem. This technology centers around a convolutional neural network (CNN) and constructs two models: the Deep Gradient Network (DGNET) for image segmentation processing and the Grouped Pooling Convolutional Neural Network (GPCNN) for peak detection processing. The models utilize radar echo data as input for inference, employing a post-processing strategy to merge the inference results and extract high voltage lines. Engineered for airborne millimeter wave radar , this intelligent detection technology has been implemented on TX2 hardware platform with various task scheduling mechanisms. Experimental results demonstrate that the detection performance of high voltage lines is robust and exhibits excellent real-time capabilities, which can meet the demands of practical engineering applications.
Key words:  high voltage line detection  deep learning  millimeter wave radar  task scheduling

版权所有:《雷达科学与技术》编辑部 备案:XXXXXXX
主办:中国电子科技集团公司第三十八研究所 地址:安徽省合肥市高新区香樟大道199号 邮政编码:230088
电话:0551-65391270 电子邮箱:radarst@163.com
技术支持:北京勤云科技发展有限公司