您的位置:首页  > 论文页面

基于Faster R-CNN的雷达时序信号频谱图的料线检测

发表时间:2019-06-28  浏览量:396  下载量:49
全部作者: 宋光然,侯庆文,倪梓明
作者单位: 北京科技大学自动化学院
摘 要: 针对雷达信号频谱中真实料面回波分量对应时序频谱的识别问题,使用改进后的Faster R-CNN来识别料线。将Faster R-CNN检测模型与3种不同特征提取的全卷积神经网络相结合,选择出最优特征提取网络,并结合正负样本均衡理论完成区域建议候选框的尺寸设置。结果显示,相比料线分割算法,料线检测的实时性和准确率更高。
关 键 词: 人工智能;模式识别;料线时序频谱的识别;全卷积神经网络;候选框的尺寸设置
Title: Stockline detection of radar timing signal spectrogram based on Faster R-CNN
Author: SONG Guangran, HOU Qingwen, NI Ziming
Organization: School of Automation and Electrical Engineering, University of Science and Technology Beijing
Abstract: In order to identify the corresponding timing spectrum of the real surface echo component in the spectrum of the radar signal, the improved Faster R-CNN is used to identify the stocklines. The Faster R-CNN detection model is combined with three different feature extraction full convolutional neural networks to select the optimal one. And positive and negative sample equilibrium theory completes the size setting of the regional suggestion candidate box, The results show that the real-time and accuracy of the stockline detection is higher than the stockline segmentation algorithm.
Key words: artificial intelligence; pattern recognition; identification of the timing spectrum of the stockline; full convolutional neural network; candidate box size setting
发表期数: 2019年6月第3期
引用格式: 宋光然,侯庆文,倪梓明. 基于Faster R-CNN的雷达时序信号频谱图的料线检测[J]. 中国科技论文在线精品论文,2019,12(3):383-389.
 
2 评论数 0
暂无评论
友情链接