华东师范大学学报(自然科学版) ›› 2017, Vol. ›› Issue (3): 120-132.doi: 10.3969/j.issn.1000-5641.2017.03.014

• 物理学与电子学 • 上一篇    下一篇

基于排序的自动剔除Switching-CFAR检测器

刘贵如1, 王陆林2, 邹姗1   

  1. 1．安徽工程大学 计算机与信息学院, 安徽 芜湖 241000;
    2. 奇瑞汽车股份有限公司 前瞻技术研究院, 安徽 芜湖 241006
  • 收稿日期:2016-09-14 出版日期:2017-05-25 发布日期:2017-05-18
  • 作者简介:刘贵如,女,硕士,讲师,研究方向为信号处理、车辆主动安全和多传感器融合.E-mail:liuguiru_yunnan@163.com
  • 基金资助:
    国家自然科学基金（91120307）；安徽省自然科学基金（TSKJ2015B12）；安徽工程大学计算机应用技术重点实验室开放基金（JSJKF201514）

Automatic censoring switching-CFAR detector based on sorting

LIU Gui-ru1, WANG Lu-lin2, ZOU Shan1   

  1. 1. College of Computer and Information Science, Anhui Polytechnic University, Wuhu Anhui 241000, China;
    2. Prospective Technology Research Institute, Chery Automobile Co., Ltd, Wuhu Anhui 241006, China
  • Received:2016-09-14 Online:2017-05-25 Published:2017-05-18

RichHTML

51

PDF

824

摘要： 针对传统恒虚警（Constant False-Alarm Rate，CFAR）检测器在非均匀噪声环境下检测性能较差的问题，本文提出了一种基于排序的自动剔除Switching-CFAR（AutomaticCensoring Switching-CFAR Detector Based on Sorting，ACS-CFAR）检测器.选择参考窗中间单元为测试单元，其余单元按照幅值升序排列，根据两个分界点位置参数，选择合适的参考单元集进行背景噪声功率估计以及结合参考单元数和目标恒虚警率计算相关系数，得到最优检测门限.经过仿真对比，ACS-CFAR检测器在均匀噪声环境下检测率为98.73%，接近于单元平均恒虚警（CA-CFAR）检测器；在非均匀噪声环境下检测率为98.16%，优于可变索引恒虚警（VI-CFAR）和自动删除平均恒虚警（ACCA-CFAR）检测器，虚警率误差均控制在0.10%以内.结果表明，本文提出的ACS-CFAR检测器在均匀噪声环境以及杂波和多目标干扰环境下均具有较好的检测性能.

关键词: 目标检测, 恒虚警, 自动剔除, Switching-CFAR, 非均匀噪声

Abstract: Because the conventional CFAR (Constant False-Alarm Rate) detectors have poor detection performance in non-homogeneous environments, an automatic censoring switching-CFAR detector based on sorting (ACS-CFAR) is proposed. The middle cell of the reference window acts as a cell under test; other cells are sorted into the ranked reference cells by ascending order according to their magnitudes. According to the location parameters of the two boundary points which can effectively discriminate between thermal noise, clutter edge or interferences plus thermal noise and interferences immersed in the clutter plus thermal noise region, the detection algorithm can effectively select a suitable cell set from the ranked reference cells to estimate the unknown background level. Combined with the number of the selecting reference cells and the desired probability of false alarm, the corresponding scaling factor can be calculated. Finally, the adaptive detection threshold will be obtained according to background noise level estimated value and the calculated scaling factor. The performances of the ACS-CFAR detector is simulated and evaluated in different simulation environments and compared to the performance of the CA-CFAR,VI-CFAR and ACCA-CFAR detectors, the detection probability of ACS-CFAR detector is up to 98.73%,98.16% in homogeneous and non-homogeneous environments, respectively. The ACS-CFAR detector performs like the CA-CFAR detector in homogeneous environments and better than the VI-CFAR and ACCA-CFAR detector in non-homogeneous environments, false alarm rate errors are controlled within ± 0.10%. The simulation results show that the proposed ACS-CFAR detector has better detection performance in homogenous and the presence of interfering targets and clutter edge environments.

Key words: target detection, constant false-alarm rate (CFAR), automatic censoring, Switching-CFAR, non-homogenous noise

中图分类号: